code/realworld/demo_realworld_perception.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test `Unseen Object Clustering (UCN)` Network"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run the shell script of `UCN Network`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# perception = \"/home/terry/Rilab/sOftrobot/UnseenObjectClustering\"\n",
    "perception = \"/home/rilab-ur/UnseenObjectClustering\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "+ set -e\n",
      "+ export PYTHONUNBUFFERED=True\n",
      "+ PYTHONUNBUFFERED=True\n",
      "+ export CUDA_VISIBLE_DEVICES=0\n",
      "+ CUDA_VISIBLE_DEVICES=0\n",
      "+ outdir=data/checkpoints\n",
      "+ ./ros/test_images_segmentation.py --gpu 0 --network seg_resnet34_8s_embedding --pretrained data/checkpoints/seg_resnet34_8s_embedding_cosine_rgbd_add_sampling_epoch_16.checkpoint.pth --pretrained_crop data/checkpoints/seg_resnet34_8s_embedding_cosine_rgbd_add_crop_sampling_epoch_16.checkpoint.pth --cfg experiments/cfgs/seg_resnet34_8s_embedding_cosine_rgbd_add_tabletop.yml\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Called with args:\n",
      "Namespace(background_name=None, cfg_file='experiments/cfgs/seg_resnet34_8s_embedding_cosine_rgbd_add_tabletop.yml', dataset_name='shapenet_scene_train', gpu_id=0, instance_id=0, network_name='seg_resnet34_8s_embedding', pretrained='data/checkpoints/seg_resnet34_8s_embedding_cosine_rgbd_add_sampling_epoch_16.checkpoint.pth', pretrained_crop='data/checkpoints/seg_resnet34_8s_embedding_cosine_rgbd_add_crop_sampling_epoch_16.checkpoint.pth', randomize=False)\n",
      "Using config:\n",
      "{'ANCHOR_RATIOS': [0.5, 1, 2],\n",
      " 'ANCHOR_SCALES': [8, 16, 32],\n",
      " 'BACKGROUND': '',\n",
      " 'CAD': '',\n",
      " 'DATA_PATH': '',\n",
      " 'EPS': 1e-14,\n",
      " 'EXP_DIR': 'tabletop_object',\n",
      " 'FEATURE_STRIDE': 16,\n",
      " 'FLIP_X': False,\n",
      " 'FLOW_HEIGHT': 512,\n",
      " 'FLOW_WIDTH': 640,\n",
      " 'GPU_ID': 0,\n",
      " 'INPUT': 'RGBD',\n",
      " 'INTRINSICS': [],\n",
      " 'MODE': 'TRAIN',\n",
      " 'NETWORK': 'VGG16',\n",
      " 'PIXEL_MEANS': array([[[102.9801, 115.9465, 122.7717]]]),\n",
      " 'POSE': '',\n",
      " 'RIG': '',\n",
      " 'RNG_SEED': 3,\n",
      " 'ROOT_DIR': '/home/rilab-ur/UnseenObjectClustering',\n",
      " 'TEST': {'ALIGN_Z_AXIS': False,\n",
      "          'BBOX_REG': True,\n",
      "          'BUILD_CODEBOOK': False,\n",
      "          'CHECK_SIZE': False,\n",
      "          'CLASSES': [0, 1, 2, 3],\n",
      "          'DET_THRESHOLD': 0.1,\n",
      "          'GAN': False,\n",
      "          'GEN_DATA': False,\n",
      "          'GLOBAL_SEARCH': False,\n",
      "          'GRID_SIZE': 256,\n",
      "          'HOUGH_INLIER_THRESHOLD': 0.9,\n",
      "          'HOUGH_LABEL_THRESHOLD': 400,\n",
      "          'HOUGH_SKIP_PIXELS': 10,\n",
      "          'HOUGH_VOTING_THRESHOLD': 10,\n",
      "          'IMS_PER_BATCH': 1,\n",
      "          'ITERNUM': 4,\n",
      "          'MEAN_SHIFT': False,\n",
      "          'NMS': 0.3,\n",
      "          'NUM_LOST': 3,\n",
      "          'NUM_SDF_ITERATIONS_INIT': 100,\n",
      "          'NUM_SDF_ITERATIONS_TRACKING': 50,\n",
      "          'POSE_CODEBOOK': False,\n",
      "          'POSE_REFINE': False,\n",
      "          'POSE_REG': False,\n",
      "          'POSE_SDF': True,\n",
      "          'RANSAC': False,\n",
      "          'ROS_CAMERA': 'D415',\n",
      "          'RPN_NMS_THRESH': 0.7,\n",
      "          'RPN_POST_NMS_TOP_N': 300,\n",
      "          'RPN_PRE_NMS_TOP_N': 6000,\n",
      "          'SCALES_BASE': [1.0],\n",
      "          'SDF_ROTATION_REG': 0.1,\n",
      "          'SDF_TRANSLATION_REG': 10.0,\n",
      "          'SEGMENTATION': True,\n",
      "          'SINGLE_FRAME': True,\n",
      "          'SYMMETRY': [0, 0, 0, 0],\n",
      "          'SYNTHESIZE': True,\n",
      "          'VERTEX_REG_2D': False,\n",
      "          'VERTEX_REG_3D': False,\n",
      "          'VISUALIZE': True},\n",
      " 'TRAIN': {'ADAPT': False,\n",
      "           'ADAPT_NUM': 400,\n",
      "           'ADAPT_RATIO': 1,\n",
      "           'ADAPT_ROOT': '',\n",
      "           'ADAPT_WEIGHT': 0.1,\n",
      "           'ADD_NOISE': True,\n",
      "           'AFFINE': False,\n",
      "           'BATCH_SIZE': 128,\n",
      "           'BBOX_INSIDE_WEIGHTS': [1.0, 1.0, 1.0, 1.0],\n",
      "           'BBOX_NORMALIZE_MEANS': [0.0, 0.0, 0.0, 0.0],\n",
      "           'BBOX_NORMALIZE_STDS': [0.1, 0.1, 0.2, 0.2],\n",
      "           'BBOX_NORMALIZE_TARGETS': True,\n",
      "           'BBOX_NORMALIZE_TARGETS_PRECOMPUTED': True,\n",
      "           'BETA': 0.999,\n",
      "           'BG_THRESH_HI': 0.5,\n",
      "           'BG_THRESH_LO': 0.1,\n",
      "           'BOOSTRAP_PIXELS': 20,\n",
      "           'BOX_W': 1.0,\n",
      "           'CHANGE_BACKGROUND': False,\n",
      "           'CHROMATIC': True,\n",
      "           'CLASSES': [0, 1, 2, 3],\n",
      "           'DISPLAY': 20,\n",
      "           'EMBEDDING_ALPHA': 0.02,\n",
      "           'EMBEDDING_CONTRASTIVE': False,\n",
      "           'EMBEDDING_DELTA': 0.5,\n",
      "           'EMBEDDING_LAMBDA_INTER': 10.0,\n",
      "           'EMBEDDING_LAMBDA_INTRA': 10.0,\n",
      "           'EMBEDDING_LOSS_WEIGHT_BACKGROUND': 1.0,\n",
      "           'EMBEDDING_LOSS_WEIGHT_MATCH': 1.0,\n",
      "           'EMBEDDING_LOSS_WEIGHT_NONMATCH': 1.0,\n",
      "           'EMBEDDING_METRIC': 'cosine',\n",
      "           'EMBEDDING_NORMALIZATION': True,\n",
      "           'EMBEDDING_PIXELWISE': False,\n",
      "           'EMBEDDING_PRETRAIN': True,\n",
      "           'EMBEDDING_PROTOTYPE': False,\n",
      "           'EMBEDDING_SAMPLING': True,\n",
      "           'EMBEDDING_SAMPLING_NUM': 1000,\n",
      "           'FG_FRACTION': 0.25,\n",
      "           'FG_THRESH': 0.5,\n",
      "           'FG_THRESH_POSE': 0.5,\n",
      "           'FREEZE_LAYERS': True,\n",
      "           'FUSION_TYPE': 'add',\n",
      "           'GAMMA': 0.1,\n",
      "           'GAN': False,\n",
      "           'GPUNUM': 1,\n",
      "           'GRID_SIZE': 256,\n",
      "           'HARD_ANGLE': 5.0,\n",
      "           'HARD_LABEL_SAMPLING': 0.0,\n",
      "           'HARD_LABEL_THRESHOLD': 0.9,\n",
      "           'HAS_RPN': True,\n",
      "           'HEATUP': 4,\n",
      "           'HOUGH_INLIER_THRESHOLD': 0.9,\n",
      "           'HOUGH_LABEL_THRESHOLD': 100,\n",
      "           'HOUGH_SKIP_PIXELS': 10,\n",
      "           'HOUGH_VOTING_THRESHOLD': 10,\n",
      "           'IMS_PER_BATCH': 16,\n",
      "           'ITERNUM': 4,\n",
      "           'ITERS': 0,\n",
      "           'LABEL_W': 1.0,\n",
      "           'LEARNING_RATE': 1e-05,\n",
      "           'MATCHING': False,\n",
      "           'MAX_ITERS_PER_EPOCH': 1000000,\n",
      "           'MILESTONES': [3],\n",
      "           'MOMENTUM': 0.9,\n",
      "           'NOISE_LEVEL': 0.05,\n",
      "           'NUM_STEPS': 5,\n",
      "           'NUM_UNITS': 64,\n",
      "           'OPTIMIZER': 'MOMENTUM',\n",
      "           'POSE_REG': False,\n",
      "           'POSE_W': 1.0,\n",
      "           'RPN_BATCHSIZE': 256,\n",
      "           'RPN_BBOX_INSIDE_WEIGHTS': [1.0, 1.0, 1.0, 1.0],\n",
      "           'RPN_CLOBBER_POSITIVES': False,\n",
      "           'RPN_FG_FRACTION': 0.5,\n",
      "           'RPN_NEGATIVE_OVERLAP': 0.3,\n",
      "           'RPN_NMS_THRESH': 0.7,\n",
      "           'RPN_POSITIVE_OVERLAP': 0.7,\n",
      "           'RPN_POSITIVE_WEIGHT': -1.0,\n",
      "           'RPN_POST_NMS_TOP_N': 2000,\n",
      "           'RPN_PRE_NMS_TOP_N': 12000,\n",
      "           'SCALES_BASE': [1.0],\n",
      "           'SEGMENTATION': True,\n",
      "           'SINGLE_FRAME': False,\n",
      "           'SLIM': True,\n",
      "           'SNAPSHOT_EPOCHS': 1,\n",
      "           'SNAPSHOT_INFIX': 'resnet34_8s_embedding_cosine_rgbd_add_sampling',\n",
      "           'SNAPSHOT_PREFIX': 'seg',\n",
      "           'SYMMETRY': [0, 0, 0, 0],\n",
      "           'SYMSIZE': 0,\n",
      "           'SYNITER': 0,\n",
      "           'SYNNUM': 80000,\n",
      "           'SYNROOT': '/home/yuxiang/Projects/Deep_Pose/data/LOV/data_syn/',\n",
      "           'SYNTHESIZE': True,\n",
      "           'SYN_BACKGROUND_AFFINE': False,\n",
      "           'SYN_BACKGROUND_CONSTANT_PROB': 0.2,\n",
      "           'SYN_BACKGROUND_SPECIFIC': False,\n",
      "           'SYN_BACKGROUND_SUBTRACT_MEAN': True,\n",
      "           'SYN_BOUND': 0.05,\n",
      "           'SYN_CLASS_INDEX': 1,\n",
      "           'SYN_CROP': False,\n",
      "           'SYN_CROP_SIZE': 224,\n",
      "           'SYN_HEIGHT': 480,\n",
      "           'SYN_MAX_OBJECT': 5,\n",
      "           'SYN_MIN_OBJECT': 3,\n",
      "           'SYN_ONLINE': False,\n",
      "           'SYN_RATIO': 5,\n",
      "           'SYN_SAMPLE_DISTRACTOR': True,\n",
      "           'SYN_SAMPLE_OBJECT': True,\n",
      "           'SYN_SAMPLE_POSE': False,\n",
      "           'SYN_STD_ROTATION': 15,\n",
      "           'SYN_STD_TRANSLATION': 0.05,\n",
      "           'SYN_TABLE_PROB': 0.9,\n",
      "           'SYN_TFAR': 0.6,\n",
      "           'SYN_TNEAR': 0.2,\n",
      "           'SYN_WIDTH': 640,\n",
      "           'TRAINABLE': True,\n",
      "           'UNIFORM_POSE_INTERVAL': 15,\n",
      "           'USE_FLIPPED': False,\n",
      "           'USE_GT': False,\n",
      "           'VERTEX_REG': True,\n",
      "           'VERTEX_REG_DELTA': False,\n",
      "           'VERTEX_W': 1.0,\n",
      "           'VERTEX_W_INSIDE': 10.0,\n",
      "           'VISUALIZE': False,\n",
      "           'WEIGHT_DECAY': 0.0005,\n",
      "           'add_percentage_max': 0.4,\n",
      "           'add_percentage_min': 0.1,\n",
      "           'cut_percentage_max': 0.5,\n",
      "           'cut_percentage_min': 0.25,\n",
      "           'ellipse_gamma_base_scale': 1.0,\n",
      "           'ellipse_gamma_base_shape': 1.0,\n",
      "           'ellipse_size_percentage': 0.1,\n",
      "           'label_dilation_alpha': 1.0,\n",
      "           'label_dilation_beta': 19.0,\n",
      "           'max_augmentation_tries': 10,\n",
      "           'max_padding_percentage': 0.5,\n",
      "           'min_padding_percentage': 0.05,\n",
      "           'morphology_max_iters': 3,\n",
      "           'num_ellipses_mean': 3,\n",
      "           'padding_alpha': 1.0,\n",
      "           'padding_beta': 4.0,\n",
      "           'rate_of_ellipses': 0.8,\n",
      "           'rate_of_label_adding': 0.5,\n",
      "           'rate_of_label_cutting': 0.3,\n",
      "           'rate_of_morphological_transform': 0.9,\n",
      "           'rate_of_rotation': 0.7,\n",
      "           'rate_of_translation': 0.7,\n",
      "           'rotation_angle_max': 10,\n",
      "           'translation_alpha': 1.0,\n",
      "           'translation_beta': 19.0,\n",
      "           'translation_percentage_min': 0.05},\n",
      " 'USE_GPU_NMS': True,\n",
      " 'gpu_id': 0,\n",
      " 'instance_id': 0}\n",
      "GPU device 0\n",
      "=> using pre-trained network 'data/checkpoints/seg_resnet34_8s_embedding_cosine_rgbd_add_sampling_epoch_16.checkpoint.pth'\n",
      "model keys\n",
      "=================================================\n",
      "conv1.weight\n",
      "bn1.weight\n",
      "bn1.bias\n",
      "bn1.running_mean\n",
      "bn1.running_var\n",
      "bn1.num_batches_tracked\n",
      "layer1.0.conv1.weight\n",
      "layer1.0.bn1.weight\n",
      "layer1.0.bn1.bias\n",
      "layer1.0.bn1.running_mean\n",
      "layer1.0.bn1.running_var\n",
      "layer1.0.bn1.num_batches_tracked\n",
      "layer1.0.conv2.weight\n",
      "layer1.0.bn2.weight\n",
      "layer1.0.bn2.bias\n",
      "layer1.0.bn2.running_mean\n",
      "layer1.0.bn2.running_var\n",
      "layer1.0.bn2.num_batches_tracked\n",
      "layer1.1.conv1.weight\n",
      "layer1.1.bn1.weight\n",
      "layer1.1.bn1.bias\n",
      "layer1.1.bn1.running_mean\n",
      "layer1.1.bn1.running_var\n",
      "layer1.1.bn1.num_batches_tracked\n",
      "layer1.1.conv2.weight\n",
      "layer1.1.bn2.weight\n",
      "layer1.1.bn2.bias\n",
      "layer1.1.bn2.running_mean\n",
      "layer1.1.bn2.running_var\n",
      "layer1.1.bn2.num_batches_tracked\n",
      "layer1.2.conv1.weight\n",
      "layer1.2.bn1.weight\n",
      "layer1.2.bn1.bias\n",
      "layer1.2.bn1.running_mean\n",
      "layer1.2.bn1.running_var\n",
      "layer1.2.bn1.num_batches_tracked\n",
      "layer1.2.conv2.weight\n",
      "layer1.2.bn2.weight\n",
      "layer1.2.bn2.bias\n",
      "layer1.2.bn2.running_mean\n",
      "layer1.2.bn2.running_var\n",
      "layer1.2.bn2.num_batches_tracked\n",
      "layer2.0.conv1.weight\n",
      "layer2.0.bn1.weight\n",
      "layer2.0.bn1.bias\n",
      "layer2.0.bn1.running_mean\n",
      "layer2.0.bn1.running_var\n",
      "layer2.0.bn1.num_batches_tracked\n",
      "layer2.0.conv2.weight\n",
      "layer2.0.bn2.weight\n",
      "layer2.0.bn2.bias\n",
      "layer2.0.bn2.running_mean\n",
      "layer2.0.bn2.running_var\n",
      "layer2.0.bn2.num_batches_tracked\n",
      "layer2.0.downsample.0.weight\n",
      "layer2.0.downsample.1.weight\n",
      "layer2.0.downsample.1.bias\n",
      "layer2.0.downsample.1.running_mean\n",
      "layer2.0.downsample.1.running_var\n",
      "layer2.0.downsample.1.num_batches_tracked\n",
      "layer2.1.conv1.weight\n",
      "layer2.1.bn1.weight\n",
      "layer2.1.bn1.bias\n",
      "layer2.1.bn1.running_mean\n",
      "layer2.1.bn1.running_var\n",
      "layer2.1.bn1.num_batches_tracked\n",
      "layer2.1.conv2.weight\n",
      "layer2.1.bn2.weight\n",
      "layer2.1.bn2.bias\n",
      "layer2.1.bn2.running_mean\n",
      "layer2.1.bn2.running_var\n",
      "layer2.1.bn2.num_batches_tracked\n",
      "layer2.2.conv1.weight\n",
      "layer2.2.bn1.weight\n",
      "layer2.2.bn1.bias\n",
      "layer2.2.bn1.running_mean\n",
      "layer2.2.bn1.running_var\n",
      "layer2.2.bn1.num_batches_tracked\n",
      "layer2.2.conv2.weight\n",
      "layer2.2.bn2.weight\n",
      "layer2.2.bn2.bias\n",
      "layer2.2.bn2.running_mean\n",
      "layer2.2.bn2.running_var\n",
      "layer2.2.bn2.num_batches_tracked\n",
      "layer2.3.conv1.weight\n",
      "layer2.3.bn1.weight\n",
      "layer2.3.bn1.bias\n",
      "layer2.3.bn1.running_mean\n",
      "layer2.3.bn1.running_var\n",
      "layer2.3.bn1.num_batches_tracked\n",
      "layer2.3.conv2.weight\n",
      "layer2.3.bn2.weight\n",
      "layer2.3.bn2.bias\n",
      "layer2.3.bn2.running_mean\n",
      "layer2.3.bn2.running_var\n",
      "layer2.3.bn2.num_batches_tracked\n",
      "layer3.0.conv1.weight\n",
      "layer3.0.bn1.weight\n",
      "layer3.0.bn1.bias\n",
      "layer3.0.bn1.running_mean\n",
      "layer3.0.bn1.running_var\n",
      "layer3.0.bn1.num_batches_tracked\n",
      "layer3.0.conv2.weight\n",
      "layer3.0.bn2.weight\n",
      "layer3.0.bn2.bias\n",
      "layer3.0.bn2.running_mean\n",
      "layer3.0.bn2.running_var\n",
      "layer3.0.bn2.num_batches_tracked\n",
      "layer3.0.downsample.0.weight\n",
      "layer3.0.downsample.1.weight\n",
      "layer3.0.downsample.1.bias\n",
      "layer3.0.downsample.1.running_mean\n",
      "layer3.0.downsample.1.running_var\n",
      "layer3.0.downsample.1.num_batches_tracked\n",
      "layer3.1.conv1.weight\n",
      "layer3.1.bn1.weight\n",
      "layer3.1.bn1.bias\n",
      "layer3.1.bn1.running_mean\n",
      "layer3.1.bn1.running_var\n",
      "layer3.1.bn1.num_batches_tracked\n",
      "layer3.1.conv2.weight\n",
      "layer3.1.bn2.weight\n",
      "layer3.1.bn2.bias\n",
      "layer3.1.bn2.running_mean\n",
      "layer3.1.bn2.running_var\n",
      "layer3.1.bn2.num_batches_tracked\n",
      "layer3.2.conv1.weight\n",
      "layer3.2.bn1.weight\n",
      "layer3.2.bn1.bias\n",
      "layer3.2.bn1.running_mean\n",
      "layer3.2.bn1.running_var\n",
      "layer3.2.bn1.num_batches_tracked\n",
      "layer3.2.conv2.weight\n",
      "layer3.2.bn2.weight\n",
      "layer3.2.bn2.bias\n",
      "layer3.2.bn2.running_mean\n",
      "layer3.2.bn2.running_var\n",
      "layer3.2.bn2.num_batches_tracked\n",
      "layer3.3.conv1.weight\n",
      "layer3.3.bn1.weight\n",
      "layer3.3.bn1.bias\n",
      "layer3.3.bn1.running_mean\n",
      "layer3.3.bn1.running_var\n",
      "layer3.3.bn1.num_batches_tracked\n",
      "layer3.3.conv2.weight\n",
      "layer3.3.bn2.weight\n",
      "layer3.3.bn2.bias\n",
      "layer3.3.bn2.running_mean\n",
      "layer3.3.bn2.running_var\n",
      "layer3.3.bn2.num_batches_tracked\n",
      "layer3.4.conv1.weight\n",
      "layer3.4.bn1.weight\n",
      "layer3.4.bn1.bias\n",
      "layer3.4.bn1.running_mean\n",
      "layer3.4.bn1.running_var\n",
      "layer3.4.bn1.num_batches_tracked\n",
      "layer3.4.conv2.weight\n",
      "layer3.4.bn2.weight\n",
      "layer3.4.bn2.bias\n",
      "layer3.4.bn2.running_mean\n",
      "layer3.4.bn2.running_var\n",
      "layer3.4.bn2.num_batches_tracked\n",
      "layer3.5.conv1.weight\n",
      "layer3.5.bn1.weight\n",
      "layer3.5.bn1.bias\n",
      "layer3.5.bn1.running_mean\n",
      "layer3.5.bn1.running_var\n",
      "layer3.5.bn1.num_batches_tracked\n",
      "layer3.5.conv2.weight\n",
      "layer3.5.bn2.weight\n",
      "layer3.5.bn2.bias\n",
      "layer3.5.bn2.running_mean\n",
      "layer3.5.bn2.running_var\n",
      "layer3.5.bn2.num_batches_tracked\n",
      "layer4.0.conv1.weight\n",
      "layer4.0.bn1.weight\n",
      "layer4.0.bn1.bias\n",
      "layer4.0.bn1.running_mean\n",
      "layer4.0.bn1.running_var\n",
      "layer4.0.bn1.num_batches_tracked\n",
      "layer4.0.conv2.weight\n",
      "layer4.0.bn2.weight\n",
      "layer4.0.bn2.bias\n",
      "layer4.0.bn2.running_mean\n",
      "layer4.0.bn2.running_var\n",
      "layer4.0.bn2.num_batches_tracked\n",
      "layer4.0.downsample.0.weight\n",
      "layer4.0.downsample.1.weight\n",
      "layer4.0.downsample.1.bias\n",
      "layer4.0.downsample.1.running_mean\n",
      "layer4.0.downsample.1.running_var\n",
      "layer4.0.downsample.1.num_batches_tracked\n",
      "layer4.1.conv1.weight\n",
      "layer4.1.bn1.weight\n",
      "layer4.1.bn1.bias\n",
      "layer4.1.bn1.running_mean\n",
      "layer4.1.bn1.running_var\n",
      "layer4.1.bn1.num_batches_tracked\n",
      "layer4.1.conv2.weight\n",
      "layer4.1.bn2.weight\n",
      "layer4.1.bn2.bias\n",
      "layer4.1.bn2.running_mean\n",
      "layer4.1.bn2.running_var\n",
      "layer4.1.bn2.num_batches_tracked\n",
      "layer4.2.conv1.weight\n",
      "layer4.2.bn1.weight\n",
      "layer4.2.bn1.bias\n",
      "layer4.2.bn1.running_mean\n",
      "layer4.2.bn1.running_var\n",
      "layer4.2.bn1.num_batches_tracked\n",
      "layer4.2.conv2.weight\n",
      "layer4.2.bn2.weight\n",
      "layer4.2.bn2.bias\n",
      "layer4.2.bn2.running_mean\n",
      "layer4.2.bn2.running_var\n",
      "layer4.2.bn2.num_batches_tracked\n",
      "fc.weight\n",
      "fc.bias\n",
      "=================================================\n",
      "data keys\n",
      "=================================================\n",
      "conv1.weight\n",
      "bn1.running_mean\n",
      "bn1.running_var\n",
      "bn1.weight\n",
      "bn1.bias\n",
      "layer1.0.conv1.weight\n",
      "layer1.0.bn1.running_mean\n",
      "layer1.0.bn1.running_var\n",
      "layer1.0.bn1.weight\n",
      "layer1.0.bn1.bias\n",
      "layer1.0.conv2.weight\n",
      "layer1.0.bn2.running_mean\n",
      "layer1.0.bn2.running_var\n",
      "layer1.0.bn2.weight\n",
      "layer1.0.bn2.bias\n",
      "layer1.1.conv1.weight\n",
      "layer1.1.bn1.running_mean\n",
      "layer1.1.bn1.running_var\n",
      "layer1.1.bn1.weight\n",
      "layer1.1.bn1.bias\n",
      "layer1.1.conv2.weight\n",
      "layer1.1.bn2.running_mean\n",
      "layer1.1.bn2.running_var\n",
      "layer1.1.bn2.weight\n",
      "layer1.1.bn2.bias\n",
      "layer1.2.conv1.weight\n",
      "layer1.2.bn1.running_mean\n",
      "layer1.2.bn1.running_var\n",
      "layer1.2.bn1.weight\n",
      "layer1.2.bn1.bias\n",
      "layer1.2.conv2.weight\n",
      "layer1.2.bn2.running_mean\n",
      "layer1.2.bn2.running_var\n",
      "layer1.2.bn2.weight\n",
      "layer1.2.bn2.bias\n",
      "layer2.0.conv1.weight\n",
      "layer2.0.bn1.running_mean\n",
      "layer2.0.bn1.running_var\n",
      "layer2.0.bn1.weight\n",
      "layer2.0.bn1.bias\n",
      "layer2.0.conv2.weight\n",
      "layer2.0.bn2.running_mean\n",
      "layer2.0.bn2.running_var\n",
      "layer2.0.bn2.weight\n",
      "layer2.0.bn2.bias\n",
      "layer2.0.downsample.0.weight\n",
      "layer2.0.downsample.1.running_mean\n",
      "layer2.0.downsample.1.running_var\n",
      "layer2.0.downsample.1.weight\n",
      "layer2.0.downsample.1.bias\n",
      "layer2.1.conv1.weight\n",
      "layer2.1.bn1.running_mean\n",
      "layer2.1.bn1.running_var\n",
      "layer2.1.bn1.weight\n",
      "layer2.1.bn1.bias\n",
      "layer2.1.conv2.weight\n",
      "layer2.1.bn2.running_mean\n",
      "layer2.1.bn2.running_var\n",
      "layer2.1.bn2.weight\n",
      "layer2.1.bn2.bias\n",
      "layer2.2.conv1.weight\n",
      "layer2.2.bn1.running_mean\n",
      "layer2.2.bn1.running_var\n",
      "layer2.2.bn1.weight\n",
      "layer2.2.bn1.bias\n",
      "layer2.2.conv2.weight\n",
      "layer2.2.bn2.running_mean\n",
      "layer2.2.bn2.running_var\n",
      "layer2.2.bn2.weight\n",
      "layer2.2.bn2.bias\n",
      "layer2.3.conv1.weight\n",
      "layer2.3.bn1.running_mean\n",
      "layer2.3.bn1.running_var\n",
      "layer2.3.bn1.weight\n",
      "layer2.3.bn1.bias\n",
      "layer2.3.conv2.weight\n",
      "layer2.3.bn2.running_mean\n",
      "layer2.3.bn2.running_var\n",
      "layer2.3.bn2.weight\n",
      "layer2.3.bn2.bias\n",
      "layer3.0.conv1.weight\n",
      "layer3.0.bn1.running_mean\n",
      "layer3.0.bn1.running_var\n",
      "layer3.0.bn1.weight\n",
      "layer3.0.bn1.bias\n",
      "layer3.0.conv2.weight\n",
      "layer3.0.bn2.running_mean\n",
      "layer3.0.bn2.running_var\n",
      "layer3.0.bn2.weight\n",
      "layer3.0.bn2.bias\n",
      "layer3.0.downsample.0.weight\n",
      "layer3.0.downsample.1.running_mean\n",
      "layer3.0.downsample.1.running_var\n",
      "layer3.0.downsample.1.weight\n",
      "layer3.0.downsample.1.bias\n",
      "layer3.1.conv1.weight\n",
      "layer3.1.bn1.running_mean\n",
      "layer3.1.bn1.running_var\n",
      "layer3.1.bn1.weight\n",
      "layer3.1.bn1.bias\n",
      "layer3.1.conv2.weight\n",
      "layer3.1.bn2.running_mean\n",
      "layer3.1.bn2.running_var\n",
      "layer3.1.bn2.weight\n",
      "layer3.1.bn2.bias\n",
      "layer3.2.conv1.weight\n",
      "layer3.2.bn1.running_mean\n",
      "layer3.2.bn1.running_var\n",
      "layer3.2.bn1.weight\n",
      "layer3.2.bn1.bias\n",
      "layer3.2.conv2.weight\n",
      "layer3.2.bn2.running_mean\n",
      "layer3.2.bn2.running_var\n",
      "layer3.2.bn2.weight\n",
      "layer3.2.bn2.bias\n",
      "layer3.3.conv1.weight\n",
      "layer3.3.bn1.running_mean\n",
      "layer3.3.bn1.running_var\n",
      "layer3.3.bn1.weight\n",
      "layer3.3.bn1.bias\n",
      "layer3.3.conv2.weight\n",
      "layer3.3.bn2.running_mean\n",
      "layer3.3.bn2.running_var\n",
      "layer3.3.bn2.weight\n",
      "layer3.3.bn2.bias\n",
      "layer3.4.conv1.weight\n",
      "layer3.4.bn1.running_mean\n",
      "layer3.4.bn1.running_var\n",
      "layer3.4.bn1.weight\n",
      "layer3.4.bn1.bias\n",
      "layer3.4.conv2.weight\n",
      "layer3.4.bn2.running_mean\n",
      "layer3.4.bn2.running_var\n",
      "layer3.4.bn2.weight\n",
      "layer3.4.bn2.bias\n",
      "layer3.5.conv1.weight\n",
      "layer3.5.bn1.running_mean\n",
      "layer3.5.bn1.running_var\n",
      "layer3.5.bn1.weight\n",
      "layer3.5.bn1.bias\n",
      "layer3.5.conv2.weight\n",
      "layer3.5.bn2.running_mean\n",
      "layer3.5.bn2.running_var\n",
      "layer3.5.bn2.weight\n",
      "layer3.5.bn2.bias\n",
      "layer4.0.conv1.weight\n",
      "layer4.0.bn1.running_mean\n",
      "layer4.0.bn1.running_var\n",
      "layer4.0.bn1.weight\n",
      "layer4.0.bn1.bias\n",
      "layer4.0.conv2.weight\n",
      "layer4.0.bn2.running_mean\n",
      "layer4.0.bn2.running_var\n",
      "layer4.0.bn2.weight\n",
      "layer4.0.bn2.bias\n",
      "layer4.0.downsample.0.weight\n",
      "layer4.0.downsample.1.running_mean\n",
      "layer4.0.downsample.1.running_var\n",
      "layer4.0.downsample.1.weight\n",
      "layer4.0.downsample.1.bias\n",
      "layer4.1.conv1.weight\n",
      "layer4.1.bn1.running_mean\n",
      "layer4.1.bn1.running_var\n",
      "layer4.1.bn1.weight\n",
      "layer4.1.bn1.bias\n",
      "layer4.1.conv2.weight\n",
      "layer4.1.bn2.running_mean\n",
      "layer4.1.bn2.running_var\n",
      "layer4.1.bn2.weight\n",
      "layer4.1.bn2.bias\n",
      "layer4.2.conv1.weight\n",
      "layer4.2.bn1.running_mean\n",
      "layer4.2.bn1.running_var\n",
      "layer4.2.bn1.weight\n",
      "layer4.2.bn1.bias\n",
      "layer4.2.conv2.weight\n",
      "layer4.2.bn2.running_mean\n",
      "layer4.2.bn2.running_var\n",
      "layer4.2.bn2.weight\n",
      "layer4.2.bn2.bias\n",
      "fc.weight\n",
      "fc.bias\n",
      "=================================================\n",
      "load the following keys from the pretrained model\n",
      "=================================================\n",
      "conv1.weight\n",
      "bn1.running_mean\n",
      "bn1.running_var\n",
      "bn1.weight\n",
      "bn1.bias\n",
      "layer1.0.conv1.weight\n",
      "layer1.0.bn1.running_mean\n",
      "layer1.0.bn1.running_var\n",
      "layer1.0.bn1.weight\n",
      "layer1.0.bn1.bias\n",
      "layer1.0.conv2.weight\n",
      "layer1.0.bn2.running_mean\n",
      "layer1.0.bn2.running_var\n",
      "layer1.0.bn2.weight\n",
      "layer1.0.bn2.bias\n",
      "layer1.1.conv1.weight\n",
      "layer1.1.bn1.running_mean\n",
      "layer1.1.bn1.running_var\n",
      "layer1.1.bn1.weight\n",
      "layer1.1.bn1.bias\n",
      "layer1.1.conv2.weight\n",
      "layer1.1.bn2.running_mean\n",
      "layer1.1.bn2.running_var\n",
      "layer1.1.bn2.weight\n",
      "layer1.1.bn2.bias\n",
      "layer1.2.conv1.weight\n",
      "layer1.2.bn1.running_mean\n",
      "layer1.2.bn1.running_var\n",
      "layer1.2.bn1.weight\n",
      "layer1.2.bn1.bias\n",
      "layer1.2.conv2.weight\n",
      "layer1.2.bn2.running_mean\n",
      "layer1.2.bn2.running_var\n",
      "layer1.2.bn2.weight\n",
      "layer1.2.bn2.bias\n",
      "layer2.0.conv1.weight\n",
      "layer2.0.bn1.running_mean\n",
      "layer2.0.bn1.running_var\n",
      "layer2.0.bn1.weight\n",
      "layer2.0.bn1.bias\n",
      "layer2.0.conv2.weight\n",
      "layer2.0.bn2.running_mean\n",
      "layer2.0.bn2.running_var\n",
      "layer2.0.bn2.weight\n",
      "layer2.0.bn2.bias\n",
      "layer2.0.downsample.0.weight\n",
      "layer2.0.downsample.1.running_mean\n",
      "layer2.0.downsample.1.running_var\n",
      "layer2.0.downsample.1.weight\n",
      "layer2.0.downsample.1.bias\n",
      "layer2.1.conv1.weight\n",
      "layer2.1.bn1.running_mean\n",
      "layer2.1.bn1.running_var\n",
      "layer2.1.bn1.weight\n",
      "layer2.1.bn1.bias\n",
      "layer2.1.conv2.weight\n",
      "layer2.1.bn2.running_mean\n",
      "layer2.1.bn2.running_var\n",
      "layer2.1.bn2.weight\n",
      "layer2.1.bn2.bias\n",
      "layer2.2.conv1.weight\n",
      "layer2.2.bn1.running_mean\n",
      "layer2.2.bn1.running_var\n",
      "layer2.2.bn1.weight\n",
      "layer2.2.bn1.bias\n",
      "layer2.2.conv2.weight\n",
      "layer2.2.bn2.running_mean\n",
      "layer2.2.bn2.running_var\n",
      "layer2.2.bn2.weight\n",
      "layer2.2.bn2.bias\n",
      "layer2.3.conv1.weight\n",
      "layer2.3.bn1.running_mean\n",
      "layer2.3.bn1.running_var\n",
      "layer2.3.bn1.weight\n",
      "layer2.3.bn1.bias\n",
      "layer2.3.conv2.weight\n",
      "layer2.3.bn2.running_mean\n",
      "layer2.3.bn2.running_var\n",
      "layer2.3.bn2.weight\n",
      "layer2.3.bn2.bias\n",
      "layer3.0.conv1.weight\n",
      "layer3.0.bn1.running_mean\n",
      "layer3.0.bn1.running_var\n",
      "layer3.0.bn1.weight\n",
      "layer3.0.bn1.bias\n",
      "layer3.0.conv2.weight\n",
      "layer3.0.bn2.running_mean\n",
      "layer3.0.bn2.running_var\n",
      "layer3.0.bn2.weight\n",
      "layer3.0.bn2.bias\n",
      "layer3.0.downsample.0.weight\n",
      "layer3.0.downsample.1.running_mean\n",
      "layer3.0.downsample.1.running_var\n",
      "layer3.0.downsample.1.weight\n",
      "layer3.0.downsample.1.bias\n",
      "layer3.1.conv1.weight\n",
      "layer3.1.bn1.running_mean\n",
      "layer3.1.bn1.running_var\n",
      "layer3.1.bn1.weight\n",
      "layer3.1.bn1.bias\n",
      "layer3.1.conv2.weight\n",
      "layer3.1.bn2.running_mean\n",
      "layer3.1.bn2.running_var\n",
      "layer3.1.bn2.weight\n",
      "layer3.1.bn2.bias\n",
      "layer3.2.conv1.weight\n",
      "layer3.2.bn1.running_mean\n",
      "layer3.2.bn1.running_var\n",
      "layer3.2.bn1.weight\n",
      "layer3.2.bn1.bias\n",
      "layer3.2.conv2.weight\n",
      "layer3.2.bn2.running_mean\n",
      "layer3.2.bn2.running_var\n",
      "layer3.2.bn2.weight\n",
      "layer3.2.bn2.bias\n",
      "layer3.3.conv1.weight\n",
      "layer3.3.bn1.running_mean\n",
      "layer3.3.bn1.running_var\n",
      "layer3.3.bn1.weight\n",
      "layer3.3.bn1.bias\n",
      "layer3.3.conv2.weight\n",
      "layer3.3.bn2.running_mean\n",
      "layer3.3.bn2.running_var\n",
      "layer3.3.bn2.weight\n",
      "layer3.3.bn2.bias\n",
      "layer3.4.conv1.weight\n",
      "layer3.4.bn1.running_mean\n",
      "layer3.4.bn1.running_var\n",
      "layer3.4.bn1.weight\n",
      "layer3.4.bn1.bias\n",
      "layer3.4.conv2.weight\n",
      "layer3.4.bn2.running_mean\n",
      "layer3.4.bn2.running_var\n",
      "layer3.4.bn2.weight\n",
      "layer3.4.bn2.bias\n",
      "layer3.5.conv1.weight\n",
      "layer3.5.bn1.running_mean\n",
      "layer3.5.bn1.running_var\n",
      "layer3.5.bn1.weight\n",
      "layer3.5.bn1.bias\n",
      "layer3.5.conv2.weight\n",
      "layer3.5.bn2.running_mean\n",
      "layer3.5.bn2.running_var\n",
      "layer3.5.bn2.weight\n",
      "layer3.5.bn2.bias\n",
      "layer4.0.conv1.weight\n",
      "layer4.0.bn1.running_mean\n",
      "layer4.0.bn1.running_var\n",
      "layer4.0.bn1.weight\n",
      "layer4.0.bn1.bias\n",
      "layer4.0.conv2.weight\n",
      "layer4.0.bn2.running_mean\n",
      "layer4.0.bn2.running_var\n",
      "layer4.0.bn2.weight\n",
      "layer4.0.bn2.bias\n",
      "layer4.0.downsample.0.weight\n",
      "layer4.0.downsample.1.running_mean\n",
      "layer4.0.downsample.1.running_var\n",
      "layer4.0.downsample.1.weight\n",
      "layer4.0.downsample.1.bias\n",
      "layer4.1.conv1.weight\n",
      "layer4.1.bn1.running_mean\n",
      "layer4.1.bn1.running_var\n",
      "layer4.1.bn1.weight\n",
      "layer4.1.bn1.bias\n",
      "layer4.1.conv2.weight\n",
      "layer4.1.bn2.running_mean\n",
      "layer4.1.bn2.running_var\n",
      "layer4.1.bn2.weight\n",
      "layer4.1.bn2.bias\n",
      "layer4.2.conv1.weight\n",
      "layer4.2.bn1.running_mean\n",
      "layer4.2.bn1.running_var\n",
      "layer4.2.bn1.weight\n",
      "layer4.2.bn1.bias\n",
      "layer4.2.conv2.weight\n",
      "layer4.2.bn2.running_mean\n",
      "layer4.2.bn2.running_var\n",
      "layer4.2.bn2.weight\n",
      "layer4.2.bn2.bias\n",
      "fc.weight\n",
      "fc.bias\n",
      "=================================================\n",
      "=============no pretrained weights===============\n",
      "model keys\n",
      "=================================================\n",
      "fcn.resnet34_8s.conv1.weight\n",
      "fcn.resnet34_8s.bn1.weight\n",
      "fcn.resnet34_8s.bn1.bias\n",
      "fcn.resnet34_8s.bn1.running_mean\n",
      "fcn.resnet34_8s.bn1.running_var\n",
      "fcn.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.fc.weight\n",
      "fcn.resnet34_8s.fc.bias\n",
      "fcn_depth.resnet34_8s.conv1.weight\n",
      "fcn_depth.resnet34_8s.bn1.weight\n",
      "fcn_depth.resnet34_8s.bn1.bias\n",
      "fcn_depth.resnet34_8s.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.bn1.running_var\n",
      "fcn_depth.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.fc.weight\n",
      "fcn_depth.resnet34_8s.fc.bias\n",
      "=================================================\n",
      "data keys\n",
      "=================================================\n",
      "fcn.resnet34_8s.conv1.weight\n",
      "fcn.resnet34_8s.bn1.weight\n",
      "fcn.resnet34_8s.bn1.bias\n",
      "fcn.resnet34_8s.bn1.running_mean\n",
      "fcn.resnet34_8s.bn1.running_var\n",
      "fcn.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.fc.weight\n",
      "fcn.resnet34_8s.fc.bias\n",
      "fcn_depth.resnet34_8s.conv1.weight\n",
      "fcn_depth.resnet34_8s.bn1.weight\n",
      "fcn_depth.resnet34_8s.bn1.bias\n",
      "fcn_depth.resnet34_8s.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.bn1.running_var\n",
      "fcn_depth.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.fc.weight\n",
      "fcn_depth.resnet34_8s.fc.bias\n",
      "=================================================\n",
      "load the following keys from the pretrained model\n",
      "=================================================\n",
      "fcn.resnet34_8s.conv1.weight\n",
      "fcn.resnet34_8s.bn1.weight\n",
      "fcn.resnet34_8s.bn1.bias\n",
      "fcn.resnet34_8s.bn1.running_mean\n",
      "fcn.resnet34_8s.bn1.running_var\n",
      "fcn.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.fc.weight\n",
      "fcn.resnet34_8s.fc.bias\n",
      "fcn_depth.resnet34_8s.conv1.weight\n",
      "fcn_depth.resnet34_8s.bn1.weight\n",
      "fcn_depth.resnet34_8s.bn1.bias\n",
      "fcn_depth.resnet34_8s.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.bn1.running_var\n",
      "fcn_depth.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.fc.weight\n",
      "fcn_depth.resnet34_8s.fc.bias\n",
      "=================================================\n",
      "model keys\n",
      "=================================================\n",
      "conv1.weight\n",
      "bn1.weight\n",
      "bn1.bias\n",
      "bn1.running_mean\n",
      "bn1.running_var\n",
      "bn1.num_batches_tracked\n",
      "layer1.0.conv1.weight\n",
      "layer1.0.bn1.weight\n",
      "layer1.0.bn1.bias\n",
      "layer1.0.bn1.running_mean\n",
      "layer1.0.bn1.running_var\n",
      "layer1.0.bn1.num_batches_tracked\n",
      "layer1.0.conv2.weight\n",
      "layer1.0.bn2.weight\n",
      "layer1.0.bn2.bias\n",
      "layer1.0.bn2.running_mean\n",
      "layer1.0.bn2.running_var\n",
      "layer1.0.bn2.num_batches_tracked\n",
      "layer1.1.conv1.weight\n",
      "layer1.1.bn1.weight\n",
      "layer1.1.bn1.bias\n",
      "layer1.1.bn1.running_mean\n",
      "layer1.1.bn1.running_var\n",
      "layer1.1.bn1.num_batches_tracked\n",
      "layer1.1.conv2.weight\n",
      "layer1.1.bn2.weight\n",
      "layer1.1.bn2.bias\n",
      "layer1.1.bn2.running_mean\n",
      "layer1.1.bn2.running_var\n",
      "layer1.1.bn2.num_batches_tracked\n",
      "layer1.2.conv1.weight\n",
      "layer1.2.bn1.weight\n",
      "layer1.2.bn1.bias\n",
      "layer1.2.bn1.running_mean\n",
      "layer1.2.bn1.running_var\n",
      "layer1.2.bn1.num_batches_tracked\n",
      "layer1.2.conv2.weight\n",
      "layer1.2.bn2.weight\n",
      "layer1.2.bn2.bias\n",
      "layer1.2.bn2.running_mean\n",
      "layer1.2.bn2.running_var\n",
      "layer1.2.bn2.num_batches_tracked\n",
      "layer2.0.conv1.weight\n",
      "layer2.0.bn1.weight\n",
      "layer2.0.bn1.bias\n",
      "layer2.0.bn1.running_mean\n",
      "layer2.0.bn1.running_var\n",
      "layer2.0.bn1.num_batches_tracked\n",
      "layer2.0.conv2.weight\n",
      "layer2.0.bn2.weight\n",
      "layer2.0.bn2.bias\n",
      "layer2.0.bn2.running_mean\n",
      "layer2.0.bn2.running_var\n",
      "layer2.0.bn2.num_batches_tracked\n",
      "layer2.0.downsample.0.weight\n",
      "layer2.0.downsample.1.weight\n",
      "layer2.0.downsample.1.bias\n",
      "layer2.0.downsample.1.running_mean\n",
      "layer2.0.downsample.1.running_var\n",
      "layer2.0.downsample.1.num_batches_tracked\n",
      "layer2.1.conv1.weight\n",
      "layer2.1.bn1.weight\n",
      "layer2.1.bn1.bias\n",
      "layer2.1.bn1.running_mean\n",
      "layer2.1.bn1.running_var\n",
      "layer2.1.bn1.num_batches_tracked\n",
      "layer2.1.conv2.weight\n",
      "layer2.1.bn2.weight\n",
      "layer2.1.bn2.bias\n",
      "layer2.1.bn2.running_mean\n",
      "layer2.1.bn2.running_var\n",
      "layer2.1.bn2.num_batches_tracked\n",
      "layer2.2.conv1.weight\n",
      "layer2.2.bn1.weight\n",
      "layer2.2.bn1.bias\n",
      "layer2.2.bn1.running_mean\n",
      "layer2.2.bn1.running_var\n",
      "layer2.2.bn1.num_batches_tracked\n",
      "layer2.2.conv2.weight\n",
      "layer2.2.bn2.weight\n",
      "layer2.2.bn2.bias\n",
      "layer2.2.bn2.running_mean\n",
      "layer2.2.bn2.running_var\n",
      "layer2.2.bn2.num_batches_tracked\n",
      "layer2.3.conv1.weight\n",
      "layer2.3.bn1.weight\n",
      "layer2.3.bn1.bias\n",
      "layer2.3.bn1.running_mean\n",
      "layer2.3.bn1.running_var\n",
      "layer2.3.bn1.num_batches_tracked\n",
      "layer2.3.conv2.weight\n",
      "layer2.3.bn2.weight\n",
      "layer2.3.bn2.bias\n",
      "layer2.3.bn2.running_mean\n",
      "layer2.3.bn2.running_var\n",
      "layer2.3.bn2.num_batches_tracked\n",
      "layer3.0.conv1.weight\n",
      "layer3.0.bn1.weight\n",
      "layer3.0.bn1.bias\n",
      "layer3.0.bn1.running_mean\n",
      "layer3.0.bn1.running_var\n",
      "layer3.0.bn1.num_batches_tracked\n",
      "layer3.0.conv2.weight\n",
      "layer3.0.bn2.weight\n",
      "layer3.0.bn2.bias\n",
      "layer3.0.bn2.running_mean\n",
      "layer3.0.bn2.running_var\n",
      "layer3.0.bn2.num_batches_tracked\n",
      "layer3.0.downsample.0.weight\n",
      "layer3.0.downsample.1.weight\n",
      "layer3.0.downsample.1.bias\n",
      "layer3.0.downsample.1.running_mean\n",
      "layer3.0.downsample.1.running_var\n",
      "layer3.0.downsample.1.num_batches_tracked\n",
      "layer3.1.conv1.weight\n",
      "layer3.1.bn1.weight\n",
      "layer3.1.bn1.bias\n",
      "layer3.1.bn1.running_mean\n",
      "layer3.1.bn1.running_var\n",
      "layer3.1.bn1.num_batches_tracked\n",
      "layer3.1.conv2.weight\n",
      "layer3.1.bn2.weight\n",
      "layer3.1.bn2.bias\n",
      "layer3.1.bn2.running_mean\n",
      "layer3.1.bn2.running_var\n",
      "layer3.1.bn2.num_batches_tracked\n",
      "layer3.2.conv1.weight\n",
      "layer3.2.bn1.weight\n",
      "layer3.2.bn1.bias\n",
      "layer3.2.bn1.running_mean\n",
      "layer3.2.bn1.running_var\n",
      "layer3.2.bn1.num_batches_tracked\n",
      "layer3.2.conv2.weight\n",
      "layer3.2.bn2.weight\n",
      "layer3.2.bn2.bias\n",
      "layer3.2.bn2.running_mean\n",
      "layer3.2.bn2.running_var\n",
      "layer3.2.bn2.num_batches_tracked\n",
      "layer3.3.conv1.weight\n",
      "layer3.3.bn1.weight\n",
      "layer3.3.bn1.bias\n",
      "layer3.3.bn1.running_mean\n",
      "layer3.3.bn1.running_var\n",
      "layer3.3.bn1.num_batches_tracked\n",
      "layer3.3.conv2.weight\n",
      "layer3.3.bn2.weight\n",
      "layer3.3.bn2.bias\n",
      "layer3.3.bn2.running_mean\n",
      "layer3.3.bn2.running_var\n",
      "layer3.3.bn2.num_batches_tracked\n",
      "layer3.4.conv1.weight\n",
      "layer3.4.bn1.weight\n",
      "layer3.4.bn1.bias\n",
      "layer3.4.bn1.running_mean\n",
      "layer3.4.bn1.running_var\n",
      "layer3.4.bn1.num_batches_tracked\n",
      "layer3.4.conv2.weight\n",
      "layer3.4.bn2.weight\n",
      "layer3.4.bn2.bias\n",
      "layer3.4.bn2.running_mean\n",
      "layer3.4.bn2.running_var\n",
      "layer3.4.bn2.num_batches_tracked\n",
      "layer3.5.conv1.weight\n",
      "layer3.5.bn1.weight\n",
      "layer3.5.bn1.bias\n",
      "layer3.5.bn1.running_mean\n",
      "layer3.5.bn1.running_var\n",
      "layer3.5.bn1.num_batches_tracked\n",
      "layer3.5.conv2.weight\n",
      "layer3.5.bn2.weight\n",
      "layer3.5.bn2.bias\n",
      "layer3.5.bn2.running_mean\n",
      "layer3.5.bn2.running_var\n",
      "layer3.5.bn2.num_batches_tracked\n",
      "layer4.0.conv1.weight\n",
      "layer4.0.bn1.weight\n",
      "layer4.0.bn1.bias\n",
      "layer4.0.bn1.running_mean\n",
      "layer4.0.bn1.running_var\n",
      "layer4.0.bn1.num_batches_tracked\n",
      "layer4.0.conv2.weight\n",
      "layer4.0.bn2.weight\n",
      "layer4.0.bn2.bias\n",
      "layer4.0.bn2.running_mean\n",
      "layer4.0.bn2.running_var\n",
      "layer4.0.bn2.num_batches_tracked\n",
      "layer4.0.downsample.0.weight\n",
      "layer4.0.downsample.1.weight\n",
      "layer4.0.downsample.1.bias\n",
      "layer4.0.downsample.1.running_mean\n",
      "layer4.0.downsample.1.running_var\n",
      "layer4.0.downsample.1.num_batches_tracked\n",
      "layer4.1.conv1.weight\n",
      "layer4.1.bn1.weight\n",
      "layer4.1.bn1.bias\n",
      "layer4.1.bn1.running_mean\n",
      "layer4.1.bn1.running_var\n",
      "layer4.1.bn1.num_batches_tracked\n",
      "layer4.1.conv2.weight\n",
      "layer4.1.bn2.weight\n",
      "layer4.1.bn2.bias\n",
      "layer4.1.bn2.running_mean\n",
      "layer4.1.bn2.running_var\n",
      "layer4.1.bn2.num_batches_tracked\n",
      "layer4.2.conv1.weight\n",
      "layer4.2.bn1.weight\n",
      "layer4.2.bn1.bias\n",
      "layer4.2.bn1.running_mean\n",
      "layer4.2.bn1.running_var\n",
      "layer4.2.bn1.num_batches_tracked\n",
      "layer4.2.conv2.weight\n",
      "layer4.2.bn2.weight\n",
      "layer4.2.bn2.bias\n",
      "layer4.2.bn2.running_mean\n",
      "layer4.2.bn2.running_var\n",
      "layer4.2.bn2.num_batches_tracked\n",
      "fc.weight\n",
      "fc.bias\n",
      "=================================================\n",
      "data keys\n",
      "=================================================\n",
      "conv1.weight\n",
      "bn1.running_mean\n",
      "bn1.running_var\n",
      "bn1.weight\n",
      "bn1.bias\n",
      "layer1.0.conv1.weight\n",
      "layer1.0.bn1.running_mean\n",
      "layer1.0.bn1.running_var\n",
      "layer1.0.bn1.weight\n",
      "layer1.0.bn1.bias\n",
      "layer1.0.conv2.weight\n",
      "layer1.0.bn2.running_mean\n",
      "layer1.0.bn2.running_var\n",
      "layer1.0.bn2.weight\n",
      "layer1.0.bn2.bias\n",
      "layer1.1.conv1.weight\n",
      "layer1.1.bn1.running_mean\n",
      "layer1.1.bn1.running_var\n",
      "layer1.1.bn1.weight\n",
      "layer1.1.bn1.bias\n",
      "layer1.1.conv2.weight\n",
      "layer1.1.bn2.running_mean\n",
      "layer1.1.bn2.running_var\n",
      "layer1.1.bn2.weight\n",
      "layer1.1.bn2.bias\n",
      "layer1.2.conv1.weight\n",
      "layer1.2.bn1.running_mean\n",
      "layer1.2.bn1.running_var\n",
      "layer1.2.bn1.weight\n",
      "layer1.2.bn1.bias\n",
      "layer1.2.conv2.weight\n",
      "layer1.2.bn2.running_mean\n",
      "layer1.2.bn2.running_var\n",
      "layer1.2.bn2.weight\n",
      "layer1.2.bn2.bias\n",
      "layer2.0.conv1.weight\n",
      "layer2.0.bn1.running_mean\n",
      "layer2.0.bn1.running_var\n",
      "layer2.0.bn1.weight\n",
      "layer2.0.bn1.bias\n",
      "layer2.0.conv2.weight\n",
      "layer2.0.bn2.running_mean\n",
      "layer2.0.bn2.running_var\n",
      "layer2.0.bn2.weight\n",
      "layer2.0.bn2.bias\n",
      "layer2.0.downsample.0.weight\n",
      "layer2.0.downsample.1.running_mean\n",
      "layer2.0.downsample.1.running_var\n",
      "layer2.0.downsample.1.weight\n",
      "layer2.0.downsample.1.bias\n",
      "layer2.1.conv1.weight\n",
      "layer2.1.bn1.running_mean\n",
      "layer2.1.bn1.running_var\n",
      "layer2.1.bn1.weight\n",
      "layer2.1.bn1.bias\n",
      "layer2.1.conv2.weight\n",
      "layer2.1.bn2.running_mean\n",
      "layer2.1.bn2.running_var\n",
      "layer2.1.bn2.weight\n",
      "layer2.1.bn2.bias\n",
      "layer2.2.conv1.weight\n",
      "layer2.2.bn1.running_mean\n",
      "layer2.2.bn1.running_var\n",
      "layer2.2.bn1.weight\n",
      "layer2.2.bn1.bias\n",
      "layer2.2.conv2.weight\n",
      "layer2.2.bn2.running_mean\n",
      "layer2.2.bn2.running_var\n",
      "layer2.2.bn2.weight\n",
      "layer2.2.bn2.bias\n",
      "layer2.3.conv1.weight\n",
      "layer2.3.bn1.running_mean\n",
      "layer2.3.bn1.running_var\n",
      "layer2.3.bn1.weight\n",
      "layer2.3.bn1.bias\n",
      "layer2.3.conv2.weight\n",
      "layer2.3.bn2.running_mean\n",
      "layer2.3.bn2.running_var\n",
      "layer2.3.bn2.weight\n",
      "layer2.3.bn2.bias\n",
      "layer3.0.conv1.weight\n",
      "layer3.0.bn1.running_mean\n",
      "layer3.0.bn1.running_var\n",
      "layer3.0.bn1.weight\n",
      "layer3.0.bn1.bias\n",
      "layer3.0.conv2.weight\n",
      "layer3.0.bn2.running_mean\n",
      "layer3.0.bn2.running_var\n",
      "layer3.0.bn2.weight\n",
      "layer3.0.bn2.bias\n",
      "layer3.0.downsample.0.weight\n",
      "layer3.0.downsample.1.running_mean\n",
      "layer3.0.downsample.1.running_var\n",
      "layer3.0.downsample.1.weight\n",
      "layer3.0.downsample.1.bias\n",
      "layer3.1.conv1.weight\n",
      "layer3.1.bn1.running_mean\n",
      "layer3.1.bn1.running_var\n",
      "layer3.1.bn1.weight\n",
      "layer3.1.bn1.bias\n",
      "layer3.1.conv2.weight\n",
      "layer3.1.bn2.running_mean\n",
      "layer3.1.bn2.running_var\n",
      "layer3.1.bn2.weight\n",
      "layer3.1.bn2.bias\n",
      "layer3.2.conv1.weight\n",
      "layer3.2.bn1.running_mean\n",
      "layer3.2.bn1.running_var\n",
      "layer3.2.bn1.weight\n",
      "layer3.2.bn1.bias\n",
      "layer3.2.conv2.weight\n",
      "layer3.2.bn2.running_mean\n",
      "layer3.2.bn2.running_var\n",
      "layer3.2.bn2.weight\n",
      "layer3.2.bn2.bias\n",
      "layer3.3.conv1.weight\n",
      "layer3.3.bn1.running_mean\n",
      "layer3.3.bn1.running_var\n",
      "layer3.3.bn1.weight\n",
      "layer3.3.bn1.bias\n",
      "layer3.3.conv2.weight\n",
      "layer3.3.bn2.running_mean\n",
      "layer3.3.bn2.running_var\n",
      "layer3.3.bn2.weight\n",
      "layer3.3.bn2.bias\n",
      "layer3.4.conv1.weight\n",
      "layer3.4.bn1.running_mean\n",
      "layer3.4.bn1.running_var\n",
      "layer3.4.bn1.weight\n",
      "layer3.4.bn1.bias\n",
      "layer3.4.conv2.weight\n",
      "layer3.4.bn2.running_mean\n",
      "layer3.4.bn2.running_var\n",
      "layer3.4.bn2.weight\n",
      "layer3.4.bn2.bias\n",
      "layer3.5.conv1.weight\n",
      "layer3.5.bn1.running_mean\n",
      "layer3.5.bn1.running_var\n",
      "layer3.5.bn1.weight\n",
      "layer3.5.bn1.bias\n",
      "layer3.5.conv2.weight\n",
      "layer3.5.bn2.running_mean\n",
      "layer3.5.bn2.running_var\n",
      "layer3.5.bn2.weight\n",
      "layer3.5.bn2.bias\n",
      "layer4.0.conv1.weight\n",
      "layer4.0.bn1.running_mean\n",
      "layer4.0.bn1.running_var\n",
      "layer4.0.bn1.weight\n",
      "layer4.0.bn1.bias\n",
      "layer4.0.conv2.weight\n",
      "layer4.0.bn2.running_mean\n",
      "layer4.0.bn2.running_var\n",
      "layer4.0.bn2.weight\n",
      "layer4.0.bn2.bias\n",
      "layer4.0.downsample.0.weight\n",
      "layer4.0.downsample.1.running_mean\n",
      "layer4.0.downsample.1.running_var\n",
      "layer4.0.downsample.1.weight\n",
      "layer4.0.downsample.1.bias\n",
      "layer4.1.conv1.weight\n",
      "layer4.1.bn1.running_mean\n",
      "layer4.1.bn1.running_var\n",
      "layer4.1.bn1.weight\n",
      "layer4.1.bn1.bias\n",
      "layer4.1.conv2.weight\n",
      "layer4.1.bn2.running_mean\n",
      "layer4.1.bn2.running_var\n",
      "layer4.1.bn2.weight\n",
      "layer4.1.bn2.bias\n",
      "layer4.2.conv1.weight\n",
      "layer4.2.bn1.running_mean\n",
      "layer4.2.bn1.running_var\n",
      "layer4.2.bn1.weight\n",
      "layer4.2.bn1.bias\n",
      "layer4.2.conv2.weight\n",
      "layer4.2.bn2.running_mean\n",
      "layer4.2.bn2.running_var\n",
      "layer4.2.bn2.weight\n",
      "layer4.2.bn2.bias\n",
      "fc.weight\n",
      "fc.bias\n",
      "=================================================\n",
      "load the following keys from the pretrained model\n",
      "=================================================\n",
      "conv1.weight\n",
      "bn1.running_mean\n",
      "bn1.running_var\n",
      "bn1.weight\n",
      "bn1.bias\n",
      "layer1.0.conv1.weight\n",
      "layer1.0.bn1.running_mean\n",
      "layer1.0.bn1.running_var\n",
      "layer1.0.bn1.weight\n",
      "layer1.0.bn1.bias\n",
      "layer1.0.conv2.weight\n",
      "layer1.0.bn2.running_mean\n",
      "layer1.0.bn2.running_var\n",
      "layer1.0.bn2.weight\n",
      "layer1.0.bn2.bias\n",
      "layer1.1.conv1.weight\n",
      "layer1.1.bn1.running_mean\n",
      "layer1.1.bn1.running_var\n",
      "layer1.1.bn1.weight\n",
      "layer1.1.bn1.bias\n",
      "layer1.1.conv2.weight\n",
      "layer1.1.bn2.running_mean\n",
      "layer1.1.bn2.running_var\n",
      "layer1.1.bn2.weight\n",
      "layer1.1.bn2.bias\n",
      "layer1.2.conv1.weight\n",
      "layer1.2.bn1.running_mean\n",
      "layer1.2.bn1.running_var\n",
      "layer1.2.bn1.weight\n",
      "layer1.2.bn1.bias\n",
      "layer1.2.conv2.weight\n",
      "layer1.2.bn2.running_mean\n",
      "layer1.2.bn2.running_var\n",
      "layer1.2.bn2.weight\n",
      "layer1.2.bn2.bias\n",
      "layer2.0.conv1.weight\n",
      "layer2.0.bn1.running_mean\n",
      "layer2.0.bn1.running_var\n",
      "layer2.0.bn1.weight\n",
      "layer2.0.bn1.bias\n",
      "layer2.0.conv2.weight\n",
      "layer2.0.bn2.running_mean\n",
      "layer2.0.bn2.running_var\n",
      "layer2.0.bn2.weight\n",
      "layer2.0.bn2.bias\n",
      "layer2.0.downsample.0.weight\n",
      "layer2.0.downsample.1.running_mean\n",
      "layer2.0.downsample.1.running_var\n",
      "layer2.0.downsample.1.weight\n",
      "layer2.0.downsample.1.bias\n",
      "layer2.1.conv1.weight\n",
      "layer2.1.bn1.running_mean\n",
      "layer2.1.bn1.running_var\n",
      "layer2.1.bn1.weight\n",
      "layer2.1.bn1.bias\n",
      "layer2.1.conv2.weight\n",
      "layer2.1.bn2.running_mean\n",
      "layer2.1.bn2.running_var\n",
      "layer2.1.bn2.weight\n",
      "layer2.1.bn2.bias\n",
      "layer2.2.conv1.weight\n",
      "layer2.2.bn1.running_mean\n",
      "layer2.2.bn1.running_var\n",
      "layer2.2.bn1.weight\n",
      "layer2.2.bn1.bias\n",
      "layer2.2.conv2.weight\n",
      "layer2.2.bn2.running_mean\n",
      "layer2.2.bn2.running_var\n",
      "layer2.2.bn2.weight\n",
      "layer2.2.bn2.bias\n",
      "layer2.3.conv1.weight\n",
      "layer2.3.bn1.running_mean\n",
      "layer2.3.bn1.running_var\n",
      "layer2.3.bn1.weight\n",
      "layer2.3.bn1.bias\n",
      "layer2.3.conv2.weight\n",
      "layer2.3.bn2.running_mean\n",
      "layer2.3.bn2.running_var\n",
      "layer2.3.bn2.weight\n",
      "layer2.3.bn2.bias\n",
      "layer3.0.conv1.weight\n",
      "layer3.0.bn1.running_mean\n",
      "layer3.0.bn1.running_var\n",
      "layer3.0.bn1.weight\n",
      "layer3.0.bn1.bias\n",
      "layer3.0.conv2.weight\n",
      "layer3.0.bn2.running_mean\n",
      "layer3.0.bn2.running_var\n",
      "layer3.0.bn2.weight\n",
      "layer3.0.bn2.bias\n",
      "layer3.0.downsample.0.weight\n",
      "layer3.0.downsample.1.running_mean\n",
      "layer3.0.downsample.1.running_var\n",
      "layer3.0.downsample.1.weight\n",
      "layer3.0.downsample.1.bias\n",
      "layer3.1.conv1.weight\n",
      "layer3.1.bn1.running_mean\n",
      "layer3.1.bn1.running_var\n",
      "layer3.1.bn1.weight\n",
      "layer3.1.bn1.bias\n",
      "layer3.1.conv2.weight\n",
      "layer3.1.bn2.running_mean\n",
      "layer3.1.bn2.running_var\n",
      "layer3.1.bn2.weight\n",
      "layer3.1.bn2.bias\n",
      "layer3.2.conv1.weight\n",
      "layer3.2.bn1.running_mean\n",
      "layer3.2.bn1.running_var\n",
      "layer3.2.bn1.weight\n",
      "layer3.2.bn1.bias\n",
      "layer3.2.conv2.weight\n",
      "layer3.2.bn2.running_mean\n",
      "layer3.2.bn2.running_var\n",
      "layer3.2.bn2.weight\n",
      "layer3.2.bn2.bias\n",
      "layer3.3.conv1.weight\n",
      "layer3.3.bn1.running_mean\n",
      "layer3.3.bn1.running_var\n",
      "layer3.3.bn1.weight\n",
      "layer3.3.bn1.bias\n",
      "layer3.3.conv2.weight\n",
      "layer3.3.bn2.running_mean\n",
      "layer3.3.bn2.running_var\n",
      "layer3.3.bn2.weight\n",
      "layer3.3.bn2.bias\n",
      "layer3.4.conv1.weight\n",
      "layer3.4.bn1.running_mean\n",
      "layer3.4.bn1.running_var\n",
      "layer3.4.bn1.weight\n",
      "layer3.4.bn1.bias\n",
      "layer3.4.conv2.weight\n",
      "layer3.4.bn2.running_mean\n",
      "layer3.4.bn2.running_var\n",
      "layer3.4.bn2.weight\n",
      "layer3.4.bn2.bias\n",
      "layer3.5.conv1.weight\n",
      "layer3.5.bn1.running_mean\n",
      "layer3.5.bn1.running_var\n",
      "layer3.5.bn1.weight\n",
      "layer3.5.bn1.bias\n",
      "layer3.5.conv2.weight\n",
      "layer3.5.bn2.running_mean\n",
      "layer3.5.bn2.running_var\n",
      "layer3.5.bn2.weight\n",
      "layer3.5.bn2.bias\n",
      "layer4.0.conv1.weight\n",
      "layer4.0.bn1.running_mean\n",
      "layer4.0.bn1.running_var\n",
      "layer4.0.bn1.weight\n",
      "layer4.0.bn1.bias\n",
      "layer4.0.conv2.weight\n",
      "layer4.0.bn2.running_mean\n",
      "layer4.0.bn2.running_var\n",
      "layer4.0.bn2.weight\n",
      "layer4.0.bn2.bias\n",
      "layer4.0.downsample.0.weight\n",
      "layer4.0.downsample.1.running_mean\n",
      "layer4.0.downsample.1.running_var\n",
      "layer4.0.downsample.1.weight\n",
      "layer4.0.downsample.1.bias\n",
      "layer4.1.conv1.weight\n",
      "layer4.1.bn1.running_mean\n",
      "layer4.1.bn1.running_var\n",
      "layer4.1.bn1.weight\n",
      "layer4.1.bn1.bias\n",
      "layer4.1.conv2.weight\n",
      "layer4.1.bn2.running_mean\n",
      "layer4.1.bn2.running_var\n",
      "layer4.1.bn2.weight\n",
      "layer4.1.bn2.bias\n",
      "layer4.2.conv1.weight\n",
      "layer4.2.bn1.running_mean\n",
      "layer4.2.bn1.running_var\n",
      "layer4.2.bn1.weight\n",
      "layer4.2.bn1.bias\n",
      "layer4.2.conv2.weight\n",
      "layer4.2.bn2.running_mean\n",
      "layer4.2.bn2.running_var\n",
      "layer4.2.bn2.weight\n",
      "layer4.2.bn2.bias\n",
      "fc.weight\n",
      "fc.bias\n",
      "=================================================\n",
      "=============no pretrained weights===============\n",
      "model keys\n",
      "=================================================\n",
      "fcn.resnet34_8s.conv1.weight\n",
      "fcn.resnet34_8s.bn1.weight\n",
      "fcn.resnet34_8s.bn1.bias\n",
      "fcn.resnet34_8s.bn1.running_mean\n",
      "fcn.resnet34_8s.bn1.running_var\n",
      "fcn.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.fc.weight\n",
      "fcn.resnet34_8s.fc.bias\n",
      "fcn_depth.resnet34_8s.conv1.weight\n",
      "fcn_depth.resnet34_8s.bn1.weight\n",
      "fcn_depth.resnet34_8s.bn1.bias\n",
      "fcn_depth.resnet34_8s.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.bn1.running_var\n",
      "fcn_depth.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.fc.weight\n",
      "fcn_depth.resnet34_8s.fc.bias\n",
      "=================================================\n",
      "data keys\n",
      "=================================================\n",
      "fcn.resnet34_8s.conv1.weight\n",
      "fcn.resnet34_8s.bn1.weight\n",
      "fcn.resnet34_8s.bn1.bias\n",
      "fcn.resnet34_8s.bn1.running_mean\n",
      "fcn.resnet34_8s.bn1.running_var\n",
      "fcn.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.fc.weight\n",
      "fcn.resnet34_8s.fc.bias\n",
      "fcn_depth.resnet34_8s.conv1.weight\n",
      "fcn_depth.resnet34_8s.bn1.weight\n",
      "fcn_depth.resnet34_8s.bn1.bias\n",
      "fcn_depth.resnet34_8s.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.bn1.running_var\n",
      "fcn_depth.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.fc.weight\n",
      "fcn_depth.resnet34_8s.fc.bias\n",
      "=================================================\n",
      "load the following keys from the pretrained model\n",
      "=================================================\n",
      "fcn.resnet34_8s.conv1.weight\n",
      "fcn.resnet34_8s.bn1.weight\n",
      "fcn.resnet34_8s.bn1.bias\n",
      "fcn.resnet34_8s.bn1.running_mean\n",
      "fcn.resnet34_8s.bn1.running_var\n",
      "fcn.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn.resnet34_8s.fc.weight\n",
      "fcn.resnet34_8s.fc.bias\n",
      "fcn_depth.resnet34_8s.conv1.weight\n",
      "fcn_depth.resnet34_8s.bn1.weight\n",
      "fcn_depth.resnet34_8s.bn1.bias\n",
      "fcn_depth.resnet34_8s.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.bn1.running_var\n",
      "fcn_depth.resnet34_8s.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer1.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer1.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer2.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer2.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.3.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.3.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.4.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.4.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer3.5.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer3.5.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.0.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.weight\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.bias\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.0.downsample.1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.1.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.1.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn1.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.layer4.2.conv2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.weight\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.bias\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_mean\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.running_var\n",
      "fcn_depth.resnet34_8s.layer4.2.bn2.num_batches_tracked\n",
      "fcn_depth.resnet34_8s.fc.weight\n",
      "fcn_depth.resnet34_8s.fc.bias\n",
      "=================================================\n",
      "[[907.66796875   0.         638.40039062]\n",
      " [  0.         907.62683105 356.03598022]\n",
      " [  0.           0.           1.        ]]\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "num_object None\n",
      "===========================================\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/rilab-ur/.local/lib/python3.8/site-packages/torch/nn/functional.py:4070: UserWarning: nn.functional.upsample_bilinear is deprecated. Use nn.functional.interpolate instead.\n",
      "  warnings.warn(\"nn.functional.upsample_bilinear is deprecated. Use nn.functional.interpolate instead.\")\n",
      "/home/rilab-ur/.local/lib/python3.8/site-packages/torch/nn/functional.py:4014: UserWarning: nn.functional.upsample_nearest is deprecated. Use nn.functional.interpolate instead.\n",
      "  warnings.warn(\"nn.functional.upsample_nearest is deprecated. Use nn.functional.interpolate instead.\")\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4 objects\n",
      "num_object 4\n",
      "# of objects detected: 4\n"
     ]
    }
   ],
   "source": [
    "CAPTURE = True\n",
    "\n",
    "import subprocess\n",
    "import os\n",
    "\n",
    "if CAPTURE:\n",
    "    if os.path.exists(f\"{perception}/label_data.npy\"):\n",
    "        os.remove(f\"{perception}/label_data.npy\")\n",
    "\n",
    "    if os.path.exists(f\"{perception}/depth_data.npy\"):\n",
    "        os.remove(f\"{perception}/depth_data.npy\")\n",
    "\n",
    "    subprocess.run([f\"cd {perception}; ./experiments/scripts/ros_seg_rgbd_add_test_segmentation_realsense.sh $GPU_ID 0\"], shell=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "label_pixel = np.load(f'{perception}/label_data.npy')\n",
    "depth_pixel = np.load(f'{perception}/depth_data.npy') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Rotation_X(rad):\n",
    "    roll = np.array([[1, \t       0, \t      0,    0],\n",
    "             \t\t [0, np.cos(rad), -np.sin(rad), 0],\n",
    "             \t\t [0, np.sin(rad),  np.cos(rad), 0],\n",
    "             \t\t [0,\t\t   0,\t      0,    0]])\n",
    "    return roll \n",
    "\n",
    "\n",
    "def Rotation_Y(rad):\n",
    "    pitch = np.array([[np.cos(rad), 0, np.sin(rad), 0],\n",
    "              \t\t  [0,\t\t    1, \t         0, 0],\n",
    "              \t\t  [-np.sin(rad),0, np.cos(rad), 0],\n",
    "              \t\t  [0, \t\t    0, \t         0, 0]])\n",
    "    return pitch\n",
    "\n",
    "\n",
    "def Rotation_Z(rad):\n",
    "    yaw = np.array([[np.cos(rad), -np.sin(rad),  0, 0],\n",
    "         \t        [np.sin(rad),  np.cos(rad),  0, 0],\n",
    "              \t\t[0, \t\t\t         0,  1, 0],\n",
    "             \t\t[0, \t\t\t         0,  0, 0]])\n",
    "    return yaw \n",
    "\n",
    "def Translation(x , y, z):\n",
    "    Position = np.array([[0, 0, 0, x],\n",
    "                         [0, 0, 0, y],\n",
    "                         [0, 0, 0, z],\n",
    "                         [0, 0, 0, 1]])\n",
    "    return Position\n",
    "\n",
    "\n",
    "def HT_matrix(Rotation, Position):\n",
    "    Homogeneous_Transform = Rotation + Position\n",
    "    return Homogeneous_Transform\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "\n",
    "# Calibration\n",
    "rotation_x   = Rotation_X(-math.pi)\n",
    "rotation_y   = Rotation_Y(-math.pi/180*55)\n",
    "rotation_mat = np.dot(rotation_x, rotation_y)\n",
    "\n",
    "# camera_p = np.load(\"planned_traj/camera_p.npy\")\n",
    "camera_p = np.array([ 0.33107886, -0.25199999,  1.08883276])\n",
    "\n",
    "position_mat = Translation(*(camera_p.tolist()))\n",
    "transform_mat= HT_matrix(rotation_mat, position_mat)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def camera_to_base(transform_mat, points):\n",
    "    ones = np.ones((len(points),1))\n",
    "    points = np.concatenate((points,ones),axis=1)\n",
    "    t_points = points.T\n",
    "    t_transformed_ponints = np.dot(transform_mat,t_points)\n",
    "    transformed_ponints = t_transformed_ponints.T\n",
    "    xyz = transformed_ponints[:,0:3]\n",
    "    return xyz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def clean_point_cloud(single_cluster):\n",
    "    # assume point_cloud is the point cloud as a numpy array with shape (n_points, 3)\n",
    "\n",
    "    single_cluster = single_cluster[:,:3]\n",
    "    # calculate the median of each dimension\n",
    "    medians = np.median(single_cluster, axis=0)\n",
    "\n",
    "    # calculate the median absolute deviation (MAD) of each dimension\n",
    "    mad = np.median(np.abs(single_cluster - medians), axis=0)\n",
    "\n",
    "    # set a threshold for outliers based on the MAD\n",
    "    threshold = 3.0 * mad\n",
    "\n",
    "    # calculate the distance of each point from the medians\n",
    "    distances = np.abs(single_cluster - medians)\n",
    "\n",
    "    # identify the outliers by comparing the distances to the threshold\n",
    "    outliers = np.any(distances > threshold, axis=1)\n",
    "\n",
    "    # remove the outliers from the point cloud\n",
    "    cleaned_point_cloud = single_cluster[~outliers]\n",
    "    return cleaned_point_cloud\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cluster_info_2_cleaned_point_cloud(label_pixel, depth_pixel, transform_mat):\n",
    "    clustering_num=len(np.unique(label_pixel))\n",
    "    # Total cluster loop\n",
    "    for cluster_idx in range(clustering_num):\n",
    "        if cluster_idx ==0: continue\n",
    "\n",
    "        # Random color \n",
    "        # color = random_color_gen()\n",
    "        cluster_xlst, cluster_ylst = np.where(label_pixel==cluster_idx)\n",
    "\n",
    "        clusters = []\n",
    "        for idx, (x,y,z) in enumerate(zip(depth_pixel[2,cluster_xlst, cluster_ylst], \\\n",
    "                                            depth_pixel[0,cluster_xlst, cluster_ylst], \\\n",
    "                                            depth_pixel[1,cluster_xlst, cluster_ylst])):\n",
    "            position=camera_to_base(transform_mat, np.array([[x,y,z]])).reshape(-1)\n",
    "            clusters.append(position)\n",
    "\n",
    "        break\n",
    "            \n",
    "    single_cluster = np.stack(clusters)\n",
    "    cleaned_point_cloud = clean_point_cloud(single_cluster)\n",
    "    return cleaned_point_cloud\n",
    "\n",
    "cleaned_point_cloud = cluster_info_2_cleaned_point_cloud(label_pixel, depth_pixel, transform_mat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def project_YZ(cleaned_point_cloud, VIZ=False):\n",
    "    # assume cleaned_point_cloud is the cleaned point cloud as a numpy array with shape (n_points, 3)\n",
    "\n",
    "    # project points onto the XY plane by ignoring the Z coordinate\n",
    "    projected_points = cleaned_point_cloud[:, [1,2]]\n",
    "\n",
    "    # compute extent of points in XY plane\n",
    "    min_x, min_y = np.min(projected_points, axis=0)\n",
    "    max_x, max_y = np.max(projected_points, axis=0)\n",
    "\n",
    "    # compute center of circle\n",
    "    center_x, center_y = (min_x + max_x) / 2, (min_y + max_y) / 2\n",
    "\n",
    "    # compute radius of circle\n",
    "    radius = max((max_x - min_x), (max_y - min_y)) / 2\n",
    "\n",
    "    # plot the circle and points\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.set_aspect('equal')\n",
    "    ax.scatter(projected_points[:, 0], projected_points[:, 1], s=1)\n",
    "    circle = plt.Circle((center_x, center_y), radius, fill=False)\n",
    "    ax.add_artist(circle)\n",
    "\n",
    "    if VIZ:\n",
    "        edge = 0.02\n",
    "        plt.xlim(center_x-radius-edge, center_x+radius+edge)\n",
    "        plt.ylim(center_y-radius-edge, center_y+radius+edge)\n",
    "        plt.axis('equal')\n",
    "        plt.show()\n",
    "    center_z = (min_y+max_y)/2\n",
    "    return center_z\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "center_z = project_YZ(cleaned_point_cloud, VIZ=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAa4AAAGdCAYAAABKG5eZAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAADQKUlEQVR4nOx9eZxT5dX/92aZJIyQMMhiRkBEQBkYtIMDVYrUglTHQbq4FndrRekiL22hipS6YFte62tVtP5oa1+sW6syFCoqvEitIjqoyCIg4mBnYRsmYMhkstzfHzfnybk392aZfXm+n4/yTHJzt+Q+5znnfM/3KKqqqpCQkJCQkOgisHX0CUhISEhISOQCabgkJCQkJLoUpOGSkJCQkOhSkIZLQkJCQqJLQRouCQkJCYkuBWm4JCQkJCS6FKThkpCQkJDoUpCGS0JCQkKiS8HR0SfQWojH46ipqUHv3r2hKEpHn46EhISERA5QVRXHjx+H3++HzZbep+o2hqumpgaDBw/u6NOQkJCQkGgBvvjiC5x66qlpt+k2hqt3794AtIvu06dPB5+NhISEhEQuOHbsGAYPHizm8nToNoaLwoN9+vSRhktCQkKiiyKbVI8kZ0hISEhIdClIwyUhISEh0aUgDZeEhISERJeCNFwSEhISEl0K0nBJSEhISHQpSMMlISEhIdGlIA2XhISEhESXgjRcEhISEhJdCtJwSUhISEh0KUjDJSEhISHRpSANl4SEhIREl4I0XBISEhISXQrScElISEhIdClIwyUhISEh0aXQLMP12GOP4bTTToPb7caECROwefPmtNs//PDDGDVqFDweDwYPHow777wTjY2N4v0lS5bg3HPPRe/evTFgwADMnDkTu3btas6pSUhISEh0c+RsuJ5//nnMnTsXixYtwpYtWzBu3DhMnz4dBw8eNN3+r3/9K+bPn49FixZh586dWL58OZ5//nn84he/ENu8+eabuOOOO7Bp0ya8/vrriEQiuOiiixAMBpt/ZRISEhIS3RKKqqpqLh+YMGECzj33XDz66KMAgHg8jsGDB+OHP/wh5s+fn7L9nDlzsHPnTqxbt0689l//9V9499138dZbb5ke49ChQxgwYADefPNNTJ48OavzOnbsGLxeLwKBgGwkKSEhIdHFkMscnpPH1dTUhMrKSkydOjW5A5sNU6dOxTvvvGP6mfPOOw+VlZUinPjZZ59hzZo1uOSSSyyPEwgEAAAFBQWW24TDYRw7dkz3n4SEhIRE94cjl40PHz6MWCyGgQMH6l4fOHAgPvnkE9PPXHPNNTh8+DAmTZoEVVURjUZx22236UKFHPF4HD/5yU9w/vnnY8yYMZbnsmTJEixevDiX05eQkJCQ6AZoc1bhhg0b8MADD+Dxxx/Hli1b8NJLL2H16tW49957Tbe/4447sG3bNjz33HNp97tgwQIEAgHx3xdffNEWpy8hISEh0cmQk8d18sknw26348CBA7rXDxw4gEGDBpl+ZuHChbj22mtxyy23AADGjh2LYDCIW2+9FXfddRdstqTtnDNnDv7xj39g48aNOPXUU9Oei8vlgsvlyuX0JSQkJCS6AXLyuPLy8lBSUqIjWsTjcaxbtw5f/epXTT9z4sQJnXECALvdDgAgXoiqqpgzZw5efvllrF+/HsOGDcvpIiQkJCQkeg5y8rgAYO7cubj++usxfvx4lJaW4uGHH0YwGMSNN94IALjuuutQWFiIJUuWAADKy8vx0EMP4ZxzzsGECRPw6aefYuHChSgvLxcG7I477sBf//pXrFy5Er1790ZdXR0AwOv1wuPxtNa1SkhISEh0A+RsuK688kocOnQI99xzD+rq6nD22Wfj1VdfFYSN/fv36zysu+++G4qi4O6770Z1dTX69++P8vJy3H///WKbZcuWAQCmTJmiO9af/vQn3HDDDc24LAkJCQmJ7oqc67g6K2Qdl4SEhETXRZvVcUlISEhISHQ0pOGSkJCQkOhSkIZLQkJCQqJLQRouCQkJCYkuBWm4JCQkJCS6FKThkpCQkJDoUpCGS0JCQkKiS0EaLgkJCQmJLgVpuCQkJCQkuhSk4ZKQkJCQ6FKQhktCQkJCoktBGi4JCQkJiS4FabgkJCQkJLoUpOGSkJCQkOhSkIZLQkJCQqJLQRouCQkJCYkuhZw7IEtISLQ+IpEI6urqcPDgQTQ1NSESiSAajYr/IpEIYrEYHA4HHA4HnE6nbux2u3HKKaegf//+ug7kEhLdEdJwSUi0IcLhMOrq6lBbW4uamhrU1tbqxvTv4cOH0RrNyO12OwYOHAi/349TTjlF/GscDxgwAA6HfPwluibkL1dCohXQ1NSEjz/+GJWVlaisrMSWLVuwb98+HDlyRLed0+nEoEGDhBGZNGmSzqgMHDgQLpdL51GRV2Wz2RCLxYQHxj2yYDCIurq6FIO4efNm1NbW4sCBAzrDaLPZMGDAAIwYMQIlJSUYP348SkpKMHLkSOmxSXR6KGprLPM6AY4dOwav14tAIIA+ffp09OlIdGMYjVRlZSW2bt2KSCQCm82G0aNHo6SkBCNGjEjxfPr169chhiEajeLgwYM6r6+6uho7d+7E+++/j3379gEATjrpJJxzzjnSmEm0O3KZw6XhkpBIg3g8jo8++gjvvfeeqZEqKipCSUmJ+G/cuHHo1atXR592zqivr8eWLVt0xvizzz4DkGrMJkyYgDPOOKODz1iiu0EaLmm4JFqAEydO4I033kBFRQX+8Y9/4MCBA7Db7cKT6upGKlscPXoUW7Zswfvvv59izEaMGIEZM2ZgxowZOO+882S+TKLFkIZLGi6JHFFbW4t//OMfWLVqFV5//XU0NjbizDPPRHl5OS699FKMHz++WxupbHH06FG89dZbWLVqFVatWoW6ujoUFBSgrKwMM2bMwEUXXSSfP4lmQRou+eBIZICqqti2bRsqKipQUVGBzZs3w2azYdKkSZgxYwbKy8sxcuTIjj7NTo14PI73338fFRUVWLVqFbZu3Qqn04mvf/3r4h4OGTKko09TootAGi5puCRMEIvFsGHDBqxcuRKrVq3C559/jpNOOgkXX3wxysvLcckll6Bfv34dfZpdFp9//jlWrVqFiooKbNiwAdFoFGeffTZmzJiBmTNn4pxzzunoU5ToxJCGSxouCYZDhw7hj3/8I5544gl8/vnnGDx4sMjPXHDBBXC5XB19it0OgUAAr776KioqKrBmzRo0NDRg/PjxuP3223HVVVfB4/F09ClKdDJIwyUNV4+HqqrYtGkTHn/8cbzwwgtQFAVXXXUVbrvtNkyYMAGKonT0KfYYRCIRvPrqq1i2bBleffVV+Hw+3HjjjZg9e7ZkJ0oISMMlDVePRTAYxF//+lc8/vjj+PDDD3H66adj9uzZuPHGG2UYsBNg7969ePLJJ7F8+XLU19dj+vTpuP3221FWVga73d7RpyfRgchlDpdVhRLdAp988gl+/OMfo7CwED/4wQ8wePBg/POf/8SePXswb948abQ6CYYPH47f/OY3+M9//oOnn34aDQ0NuOyyy3D66afj/vvvx4EDBzr6FCW6AKTHJdFlEYlEUFFRgccffxzr169H//79ccstt+DWW2/Faaed1tGnJ5ElKisr8fjjj+Ovf/0rYrEYvvOd7+D222/HpEmTZEi3B0F6XBLdGvF4HCtWrMCoUaPw3e9+F+FwGM888wy++OILPPDAA9JodTGUlJRg+fLlqKmpwa9//WtUVlZi8uTJOO+88/Dmm2929OlJdEJIwyXRZaCqKlavXo1zzjkH1157LcaNG4cPPvgAb731Fq655hrJDuzi6Nu3L+6880588skn+Oc//4loNIopU6bg4osvxocfftjRpyfRiSANl0SXwNtvv40LLrgAl156Kfr27Yt33nkHL7/8Ms4+++yOPjWJVobNZsM3v/lNbN68GS+++CI+++wznHPOObjmmmuwd+/ejj49iU4AabgkOjW2b9+Oyy67DOeffz6OHTuGNWvW4P/+7/8wceLEjj41iTaGoij47ne/i+3bt+MPf/gD3nzzTZx55pmYM2cO6urqOvr0JDoQzTJcjz32GE477TS43W5MmDABmzdvTrv9ww8/jFGjRsHj8WDw4MG488470djYKN7fuHEjysvL4ff7oSgKXnnlleaclkQ3QlVVFW644QaMHTsW27ZtwzPPPIMtW7bg4osvlgn7HgaHw4Hvf//72LNnD+677z4888wzOOOMM7Bw4UIEAoGOPj2JjoCaI5577jk1Ly9P/eMf/6hu375d/f73v6/6fD71wIEDpts/88wzqsvlUp955hl137596tq1a9VTTjlFvfPOO8U2a9asUe+66y71pZdeUgGoL7/8cq6npQYCARWAGggEcv6sROfBoUOH1J/85CdqXl6eOmDAAPXRRx9Vw+FwR5+WRCdCfX29+vOf/1x1u91qv3791P/+7/9WQ6FQR5+WRAuRyxyes+EqLS1V77jjDvF3LBZT/X6/umTJEtPt77jjDvXCCy/UvTZ37lz1/PPPNz8habh6JCKRiPrggw+qvXv3Vnv37q3ee++96vHjxzv6tCQ6Mf7zn/+ot956q2q329XBgwerL7zwQkefkkQLkMscnlOosKmpCZWVlZg6dap4zWazYerUqXjnnXdMP3PeeeehsrJShBM/++wzrFmzBpdcckmzPESJ7odt27Zh4sSJ+MUvfoGbbroJn332Ge6++26cdNJJHX1qEp0YhYWFePLJJ7Fjxw6UlJTgiiuuwOWXX46DBw929KlJtDFy6v52+PBhxGIxDBw4UPf6wIED8cknn5h+5pprrsHhw4cxadIkqKqKaDSK2267Db/4xS+af9YAwuEwwuGw+PvYsWMt2l9XwopNVVi2YS9mTxkOAGK8eV89Vm+tQVmxH49cfQ5+9OwHur87G6LRKH7zm99g8eLFGD58ON555x2UlpZ29GlJdDGMHDkSL7/8Ml544QXccccdKCoqwmOPPYYrrriio09Noo3Q5qzCDRs24IEHHsDjjz+OLVu24KWXXsLq1atx7733tmi/S5YsgdfrFf8NHjy4lc6482PZhr2obghh2Ya9uvHqrTWIqcDqrTUAIP5e9VENzn9wPVZsqsKKTVVi3JEgL2vhwoWYO3cutmzZIo2WRItwxRVXYPv27bjgggtw5ZVXSu+rGyMnw3XyySfDbren6IkdOHAAgwYNMv3MwoULce211+KWW27B2LFj8a1vfQsPPPAAlixZgng83uwTX7BgAQKBgPjviy++aPa+OiOMBob/PXvKcBT6PJg9ZbhuXOT3AkDKvw6bYmroOgLRaBQPPPAASkpKcOLECbzzzjtYsmQJ3G53h5yPRPfCgAED8Le//Q3PP/88NmzYgKKiIrzwwgsdfVqtjnQL0M6yOG1L5GS48vLyUFJSgnXr1onX4vE41q1bh69+9aumnzlx4gRsNv1hSAVabYFMosvlQp8+fXT/dScYDQz/e/O+etQFQti8r1433l9/AgBS/oUC2BWgZGhflAztK8btje3bt+OrX/2q9LIk2hzd3ftKtwDt6MVpeyDnUOHcuXPx1FNP4emnn8bOnTsxe/ZsBINB3HjjjQCA6667DgsWLBDbl5eXY9myZXjuueewb98+vP7661i4cCHKy8uFAfvyyy/x4YcfClmXffv24cMPP8T+/ftb4RK7JowGhv/NQ4LG8KAZojEVMRWorDqKyqqjYtxeiEajWLJkCb7yla8gGAxKL0uiXWDmfb344osdfVpZI9uoixHp3usuyImcAQBXXnklDh06hHvuuQd1dXU4++yz8eqrrwrCxv79+3Ue1t133w1FUXD33Xejuroa/fv3R3l5Oe6//36xzfvvv4+vf/3r4u+5c+cCAK6//nr8+c9/bu61dWkYDQz/u8jvxdbqgAgF0vi0k/OxemsNJo/sDwCYPLI/Vm+tQZHfiyPBphQyR3tg+/btuOGGG7Blyxb89Kc/xS9/+UtpsCTaFVdccQWmTJmC22+/HVdccQW++93v4rHHHsOAAQM6+tTSgntOsyYO1f397/kXYtbEoaafmzVxqOV73QWyrUknBWcOzpo4VPc3/YALfVr7c7Pxv+dfiPMfXI/qhhB8HifyXY4Uw9XWP+6nn34aP/jBD3D66afjT3/6EyZMmNCmx5OQyARiHtrtdrz88suWKY7OgBWbqrB07S4AwLzpowBA93d3M06yrUk3wKyJQy1XVVbkjNlThsPncSIYjmLFpioRXgxHY+1KzohGo/iv//ov3HDDDZg1axa2bNkijZZEp8AVV1yBbdu2YeTIkZgyZQr+9Kc/dfQppcXxxggaQhHxvBr/7qmQhquLIFuDw3/YG3cfQkwFonG13cgZDQ0NuPTSS/E///M/eOSRR/DUU0/J0KBEp8LAgQPxxhtv4Prrr8dNN92EuXPnIhqNdvRppWDZhr2IqYACIBiOYunaXYgl4mP1waZuzRrMhJxzXBIdAwoR8lAhGTE+jqkag3D2lOEirBBJ/Npf31EntqFxa2LXrl2YMWMGDh06hFdffVWnsCIh0ZmQl5eHJ598EuPGjcOPf/xjbNu2Dc8//zz69m1/tq0VSob2RV0ghDyHDQ2hCDzOpJ8RisRE7qsnQnpcXRDpQoUepw1xFdi8rx5DCnrpPheKxBGKxMW4NVds//znPzFhwgTY7XZs3rxZGi2JTg9FUXDHHXfgtddeQ2VlJUpLS7Fz586OPi0BImS5HHYU+jxwOey693uy1yUNVxdBNqHCWROHoikahwqNHr+9Jn3LB/LIWgJVVbF06VJceumlmDRpEjZt2oQzzjijxfuVkGgvXHjhhXjvvffgcrkwceJErFmzpqNPCSs2VSEYjsLncepYwmZeV0+ENFxdBNyzspJ8AvSqGTS2Qjgaa9E5NTY24vrrr8dPf/pT/OxnP8PKlSu7BaNToufh9NNPxzvvvIMpU6bg0ksvxW9+85sWCSS0FMs27EVDKIJ8lwOVVUdR3RBCZdVRFOS7AGh5L5/H2a1rtdJB5ri6IHi+C9DXZaWoZqRBONp8ya2amhp861vfwtatW/HXv/4VV199dbP3JSHRGdC7d2+8/PLLuOeee/Dzn/8cW7duxVNPPQWPx9Pu55LuGW+vcpbODOlxdRG0BY09rqJZMfL33nsP48ePR3V1Nf71r39JoyXRbWCz2XDffffhueeew0svvYTJkyejpsZalaa9ka5MpidBGq4ugmxDhZNH9odd0f6lcbpG97kawo0bN+LCCy/EkCFDhAGTkOhuuPLKK/HWW2+hrq4OX/va11BV1b4kiHTPuIQ0XF0SvBbLWHTMpaFonC5Sn0uM/PXXX8c3v/lNTJgwAevWrcMpp5zS8ouRkOik+MpXvoK33noLADB58mR8+umn7XZsK+awhAYp+dRFQPJNhT4PguEoGkIR+DxOzJs+CotWbkNMhc4j4/Hwg8caEYlrX3Ohz43qhkax3xnjsmsyuXr1anznO9/BN77xDfztb3/rkLi/hERHoLq6GhdeeCGOHz+O9evX48wzz+zoU+qWkJJP3RBWqy6qrqeiYzM47MlgITdagNZkMhNefvllfOtb38LFF1+Ml156SRotiR6FwsJCvPnmmygoKMAFF1yAjz/+uKNPqcdDGq4uCJ7HolBhb7cTgHls3OWwW+a53M70P4Hnn38el19+Ob797W/jhRdegMvlauWrkZDo/Bg0aBA2bNgAv9+PKVOm4IMPPuiwc+kJjSIzQRquLgJukHgea9bEoch3OYQ+oVlsfN70USgf54cCwOO0Y8Y4P5w2zZSNGNDb8pgvvfQSvve97+Gaa67BM888A6fT2U5XKyHR+XDyySdj/fr1GD58OKZNm9Ymnle6HlwESdaQhqvLwEjI4GHDbJK3lVVHoQJoisZQOqwA0UTO6+Nqc3WNf/zjH7jqqqtw+eWX409/+pNo+ikh0ZPRt29frF27FoMHD8bUqVPxySeftNq+V2yqwqKV24RR+tGzH+DuV7S/l67dlVUTyZ4Cabi6CLLtXLx07S7xQ+crs9lThsOuaAK7yzbsFSFCs1Dh66+/ju985zu49NJL8Ze//EUaLQkJhr59++K1115D//798Y1vfAN797aO58Pz1SVD+6KC5Z8DoUhaL6unhQ+l4eoisKrjAvShA1LDCEfjKS1M8hw2KND+phChMVT45ptv4rLLLsPUqVPx3HPPyfCghIQJ+vfvjzfeeAMnnXQSLrzwwlap86J8dZ7DlkKaUpEkYJmFCnta+FAari6IdKFCl0P7Sl0Om85LW7ZhL0IRTYC3suqoEODlQrybN29GWVkZzj//fPz9739HXl5eu1+bhERXwaBBg7Bu3TrY7XZ84xvfQF1dy1sFHW+MiOfUiMWXjcGsiUNNQ4U9LXwoDVcXQbYrKiPjkH7M/fI1I+S0K5g9ZTh6uzWZyt5uB1ZsqsL4Bc/joksuRXFxMV555RXZ/FFCIguceuqpWL9+PU6cOIFvf/vbCIfDzd4XbxzpMYTwPc5kuN5M9qmnSUFJw9VFkG2o8PUdBxKNIg9g87561AVC2LyvXnhWkZiKzfvq0RDSOr42hKJ49PUd+PjPdyMUVfDyyy8jPz+/w65TQqKr4bTTTsPLL7+MLVu2YPbs2c1WladnvHycH00GAeye3MLEDNJwdRHwFVW6UGFjRGtV0hiJYfXWGsRUrTdXWbFf7Gv11hoU+jSPyuu2Y+fzv0Hk8Bf49s8fxuv7GntUkldCojUwYcIEPPXUU/jTn/6ERx55pEX72rj7kPC8qP4yncBAT4Q0XN0MYwu94t+yYj/sClBW7EfpsAJ4nBo5o6zYj/pgBACwf8PzOLZ9A/qV3YmdTf10rEQJCYnsce211+KnP/0p5s6di9deey3j9pwJyKnwgCbf5vU4BSmD8lsSGqTh6oJIFyrk/bhKhxVgkNeD0mEFgpxhU4DSYQVojMRw4tPNaHjzaXjPuwr5Z05COBpDIBTpyEuTkOjSWLJkCaZPn44rr7wSe/bsSbutUeWGvCxA864oX11W7MesiUN7HOU9HaTh6iLgP9psC5DT1XGdZqvH4VW/hWfEBHgnXQMAiMZUwWaiduESEhLZw26349lnn8XAgQMxY8YMBALmBf6AeZcHAEIFx1i72dMo7+kgDVcXQUtZhbMmDsXiy8ag0OfBtef0w9tPzoejzwCcXDYXiqL9DEhBHtDi7BISErnD6/WioqICtbW1+N73vodYLGa6HeWy6FkLhCJi4dgvPw/1wSZRdwkgpS6zJ8PR0ScgkR14uxJuxGZNHJpi1GiVVjqsIGU/aiyGRxfOQSz0JQZe9xBsrl6mxwtHY1ixqUq2CWfg9wMwb6dO45KhfVFZdVS2W++hGDlyJJ5//nlccskluOuuu/Dggw+K91ZsqsLStbvQkAjLN4QiWPjKNl3t1vaaAGKJF8jjylY9pydAGq4uiJKhfVEXCOlWYvxvGpsZuG0vP4ovP3gHE+/4HWp6DUpzFCXl8z0dxgVCunFdICTCsvx9eR97DqZPn46lS5di7ty5GDt2LL73ve8JEkbMwJhXoeW33E4bXA47Jo/sLzwxnhLgC6SeDGm4ugisvCogdSVGY6NB8+x7E8crK3DBjfPxRf7paVsjuxw2+aAYYLVAMBsP8rpRF2g0fV+i5+AnP/kJPvroI9x8880YOXIklm04nmK0AM1ojS304kiwSTxvMlxvDZnj6iJI18rb6j1u0D788EOsX/4ATjr7m2gaMRVFfm/a4w0p0EKIwXAUS9fukkwm6BcImcZ1gUZRQ0e5DBni6XlQFAVPPPEEzjnnHHzrW9/CdV85WZAwOFRonRo4y7AhFBFEDUCSMzik4eoiaI6kCxmx758/GNdffz0Gnz4CY77zE9z+9TOw5+CXaT+7vSZg+vD0ZFgtEMzGVENHq+uepCMnoYfb7caLL76IL7/8Eiv+51c43mhecuJ22nS/I5/HCZ/HmVP7op4CGSrsgshEzqAxhbaWPbwUO3bswHd/9b/YfCyKzfvqhcKGFQZ53Zg9ZTgWV2xHJK4KrcOegHQkjFxAXaknj+wvvK1cCB6SzNF9cOqpp+KKOXfhqft/hv69xqDX8HPFewoAr8ep+50AQL7LofsNzJo4VP4eElDU5gprdTIcO3YMXq8XgUAAffr06ejTaVMY2X5Wk+GildsQqvsMtX+5E/fcfRdWNJWKfj+93Q6hV2gGuwLsXVKG4QtWi8/sXVLWXpfYoTj/wfWobgih0OcBgGaNydvS7rUTDaFIs/b37/kXttNVS7Q1zluyDh/8v58jcnAfTv/BMjQ58qEAKB/nR2XVUQTDUdPfSabfQHdh/+Yyh8tQYTfGN0f3x5HVD+HkwcNx11136SSg0hktACIHZvy3J4DXyzR3XOT3CuMVjsabvT+Jrg0uHDD+tAIM+OYPYY834fAb/w+A5mlVVh3VST2Z5bHToSfmvmSosAsi21DhgD0rETnyBX6ytAIvbKnFxt2HRPiKQ0EqwXB//Qms2FSFj6u1yv89B7/E+Q+uN/XyuvIqzwxWLM1cxkeCTSgr9mP11hoAaov2J9H1QLVaVFRMdVtK75NRcOEtqPvHw3CMOA/Fk6fitJPzURcIYfLI/njk6nN0+zE2izQLJRvZrj0B0nB1QRhp6sa/l23Yi4tPCeGeux5Dn4lXYE2NG6jZKwoetck0CatY8dK1u8R7oUhMiO8ajWV3M1y50N7TjckAAorOg8p1fxJdD0RsArSFIdcAzRv9Dbh3/Av1r/4eW089C0eC/U0XKtnUDc6aOLRHFiY3K1T42GOP4bTTToPb7caECROwefPmtNs//PDDGDVqFDweDwYPHow777wTjY2NLdqnhDXi0Qj+cN88nHr6CJx18fUpLCXe4gTQN6mzKVpeZkhBL93DZkuof4ajGqmjOzOccqG9pxtTyI97XLnuQ6Jrgp43p02BCv3iMM9uQ7/pP0Q8EoZnyzOWz1I2LFbjdj0FORuu559/HnPnzsWiRYuwZcsWjBs3DtOnT8fBgwdNt//rX/+K+fPnY9GiRdi5cyeWL1+O559/Hr/4xS+avc+ejnTq8Ms27MXOfz6N/3y2B7fevRQ2R2poEEgaokKfGyMGnAQAcNoU9HE7EVO1mhL+sMWt3LJuiFxo7+nGZIBcDnuz9yHRtRExeXAcdgVDhwzGLfMW4ZONq9D3yLYe1b24NZBzqPChhx7C97//fdx4440AgCeeeAKrV6/GH//4R8yfPz9l+7fffhvnn38+rrlGUyA/7bTTcPXVV+Pdd99t9j57OtKFCj/ftQ3vbHoBl91wB/5Z69EZOB4qpOepLtCIuoDm/cZVFfOmj8KyDXtRHwwjFImnHjzReKGrhQoz0dCN2oKtAfpeaN+5YvO++pyo81bjrvD9dCdYyToloT1DXyu7HP/Z8n+49dZbsW3bNvTtqw8NZxsq7GrPYmsgJ8PV1NSEyspKLFiwQLxms9kwdepUvPPOO6afOe+887BixQps3rwZpaWl+Oyzz7BmzRpce+21zd4nAITDYYTDYfH3sWPHcrmUbolopAl/uG8ePAOG4rIb5mDLf74UuZLSYQW4f/UONEbiKPJ7sefgcTRG4iJsaOySXJCfh+qGxpRjUKiwX34eqhtCXaa+K9MkYKUt2JIxraKJXp/rPlrrnHrKZNYZkNloAdFYHNUNITzx5md4/g9/wJgxY3DnnXfiz3/+s247s9y12bgnSrPlZLgOHz6MWCyGgQMH6l4fOHAgPvnkE9PPXHPNNTh8+DAmTZoEVVURjUZx2223iVBhc/YJaA3bFi9enMvpdxtYsQrvWXw//vPZHgy67iE89e8vACTZaY9cfY7Y7kiwCXeVjcayDXuFgjypydM2Voir0LEN6d/OjkyEi3Tagi0ZZ3Ps9jgnidaHkVn7o2c/QMVHGvHJrgB5DjtCrNCf2LvRuCpCwaeeeip+97vf4aabbsIVV1yBSy65RGxvLDhON+5pi5M2ZxVu2LABDzzwAB5//HFMmDABn376KX784x/j3nvvxcKFC5u93wULFmDu3Lni72PHjmHw4MGtccqdHmahwv+peBdb33oeZd/7PupHjcm4MsuktkEFymZU+WUb9sJhVxCJqXDYFXQFZKK4k7ZgSyjwVlT25tLrW/OcJFofxgUkZ+v2djtFdAIAigu9OO3kfBHZ4LT3G264Ac888wzmzZuHiy66CA6HJHtnQk536OSTT4bdbseBAwd0rx84cACDBpm3yFi4cCGuvfZa3HLLLQCAsWPHIhgM4tZbb8Vdd93VrH0CgMvlgsvlyuX0uzW+WPe/yHO5UX7d7Xj6/cyq0lZhiKVrdyGmQhQoG42Wkvjs/at3aIbL1jUMV6awi1X/rM376pvl6fTLzxN1bx3tcZFnYHWNfNzTVu4tAT0vwXAUKzZVocjvxdZEBKIhFNGxdfcc/BJXnDvYtE+eoij49a9/jfHjx+Mvf/kLbrrppna9jq6InAxXXl4eSkpKsG7dOsycORMAEI/HsW7dOsyZM8f0MydOnIDNpicv2u3aF6qqarP22dNhXOk99OIGHHhvDYZ+8/t4+v1DWSVxmxsPV6GRBqKJIH40XTC/EyHbsIvxtWUb9jbL06FGgGZtaNrb4yJlhmxyZtJw5YbjjRFxT4NhvRqNy2ETocLGSCwtiaKkpARXXHEFFi1ahKuvvhoej6fdrqErImefdO7cubj++usxfvx4lJaW4uGHH0YwGBSMwOuuuw6FhYVYsmQJAKC8vBwPPfQQzjnnHBEqXLhwIcrLy4UBy7RPCT2M3kPeRy8ir08/3PPzO5HncmeVxKWHaOnaXbqHb0hBLzRUB+DzOHC8MYpBXncKSWP11hqRfDaj+3YnZJsgz9aDa+99NGd/EpnBSRhUXL7qo2SocMY4v44QVT5OIz4ZvWCuRLNncBlqa1/C448/jv/6r//qqEvrEsjZcF155ZU4dOgQ7rnnHtTV1eHss8/Gq6++KsgV+/fv13lYd999NxRFwd13343q6mr0798f5eXluP/++7Pep4Qe3HuorKzEu+tWY/ny5bjpglHifQBpe2jRhBwMR8XDR8K8QDJUWGPCLCzye7Gz9hgicRXOLhIqlJBoTZAnrgDIc9iw6qMa0cXY63GidFhBipd//oPrU7xgHg2pd/TDyeMvxgMPPIBbbrkFXq+3oy6v00Oqw3dxTJs2DdXV1di6dWtKUpernP97/oW6v3mNEXVanTd9FDbvqxcPIQB4nLaUei6fx4lwNI5QJAaP046d936zPS5Vh1zbg5iNs/FAyDNtrjp8ayjMt2Qfue7PWHsma8XMYdQiBKg6Swun83tJ92rp2l0Aks8ZJ2rQ7/mqopPw8yun4M4779Qt7nsCcpnDpeHqgqAf+dd61eLXP56FUbMW4+47rk/b4sT4Hk3IPo9ThAoLfR6UDO0rKL3UTnyrgfJOHVy1BLQNBfku02O05WTWmq1HspnIczWK/fLzsL0mYFonR2MqP8hmH5m2bW1DnavR7GntVzj13QjyugCktCmxK8Diy8YI0V2fx4kPF12ke252VDyJ//mf/8Gvn9+AZ7cdb/dnq6MgDVc3N1znP7ge/zl6AkeemQfFZkPB1b/BqX17WXpVZj98WvHlOZIe1Yxxfl3+yuO0ozESEyvKQp+W7yr0uXHwWBiRuAqbotV20QNo9PLaCu3lcTV3gsjGsGa6P+11LzlyYSDycXecSK2wYlMV7n5lm+X7VELi8zhFM0gAIidW6POI3ltmz83q276C008/HXkjvwb3lFtNIybdcaEg+3F1c8yeMhzu/7yHYPUu/Gj+Ipzat5eOgMGNlpWmIVcuJ2zcfQh5juRPwtglmfJd1Q2NgpRB3IyGUAQrNlUJJY2uoqjRVmiNXluyL1fHgvfS4n9TyM8KKrSc8eSR/XWv93ZrItezpwzH5JH9ddvw59bn82HBggU49P4axBtqxPcvfw9JyEq3LohYNIr9r/0RxRMvwFlfmYgNG8wbyJkVKhtXy/VBTTaLzBfPZxldcYdNQSSuwudJdk/mBcoU8wfaXlHDqoA6l3Fb0sOzLTrOZR/tAbqvucpNdUePy0hfp795fRZ/FjjyHHZs3H0IDaGIuFcNoQioXn/j7kOIqRD5ZSPmzJmDu+/7DY5sXIHK07TnVXYNSEIari6IXz30GBoP/wfK1QstH67m0Jsnj+yvCxX6PA7kuxyobmiETQHyXQ40hCI43mjePTnYFIUtkRtpa7Jha/TMyqbAN1dxXno/m9yY2b75/vg+cgmNtoRGT/c11+LndLnVrgpOVDn/wfUoGdoXwXBU1+7HqpN4KBKDy2GDz+NEMBzF5JH9xWJg0cptusgGkGokPR4Prr1jHpYvmY+LTwnpzkeWLcgcV5dDKBRC4ZBhsBeOwe+eWA4gMwHDLEYOaKtlZ0K6CYB4yCgM6Ex4WIDGLryrbDSWrt0lpGxcDjvyXXZTMV5Ay5kZO7q2FtqLnNHcfWRjuMwYi3x/PI+Ry/V2JkZjV87FrNhUhcUV29kzoNceTAcFwL0zx+i+Yyo3ian6/JcV8eLpt/Zi9re/jjNOH4atm96U5AwG6XF1EdCPdkzwAxw9fADnXPfrjJ8xeiX8b2Kp1RgEdXlBMR83RuLYvK9etEbxOLUVo9FocWO3emtNmxmu9vK4mruPbEKZZtdg3F9zrrctr6s5464Co9doVHm3MloKAD8jXBBzkKuvcGNjZnzMhHL/8FYV8ideiY9XLcWOHTuwbENdtw7N5gJpuLoIaCLc+sz/g2/Euah3nmyaZzBOmDwmzmPkpJc2ttCLj6sDcDvtmDyyv66Gi8fv/T63TkS0MRK36NeVxCCvuzVvgQ7NFa5tLUklKtyuaWiECs1I90404Vy9tQZFfm/GSd2Y5+D7M+Yxcrne1pCK6okCv2bPTjbw+7Tf+eSR/bFx9yGEozEdgSMYjmYkdJhh9pTheCw6FU1v/RnLli3D7O/Nk6HCBKTh6kRIFwqYPWU4frtiDar+8wnu/M1TeDfmMQ1DpQtR8ffoIQ2GtbxUKBJDZdVRHSEjwOL3NQ2NupqusYVaTy+j8YoyL81MdaO10Nz6qtbKBdFE5Hba0RSNiUmOQmykVZjrpO522lGQn5cyOeVyvZ1NbqqrwHiPqdYqE2ixEQxHRU1kKBLXkTKAVKJLJq+JvLC7G3+ARx55BEuWLOnxnhZBGq5OBDP6Ojdide9WoN+gQpx93oV491+ft+hY9JAaJZ/uWblNUNzdLKbvdtqxv/6E+Pz++hNwOewphqt8nF8UZrqdrVNtYVZblI1objrJq5aCiCwjBpwk7gutuGmcaeKnWjqiRNP+jgSbdNfd0QYgl75QxrHVd9cVcLwxs9ECtN95Qb5LPEtUgMwNIIBmG/Vbb70VS5YswTPPPIMf/OAHuV1EN4UkZ3QipCNWTFi0Eu89cAUGXzgLp379e5aJcKskvvG9kqF9RUhrf/0JhKMxuBx2HGuMCMNVXOjFkWBYFB0Hw7GUHBc3XD6P01TOpqWg87YiLmT6XEeQGJp7fp2FWJHrvc50ja21v7YCF83lajKZoEBbrFVWHRXhX3qW5k3XtEO51BPQPMblzJkzsW/fPnz44YdQlO6pDyoLkLsIjAWOHLwgEQDOaHgfUOP42Y9n694zbpfte0R7314TQL7LgVAkjoZQ0mgB2nvVrOh4SEEv8Z7LYde1NPF5nKJmpbXrTajwcpDXnVMBZqYiYL6/bAqG+ZjuZVmx3/J+Z4LZd1Xk9zb7nFrjupp7rzN9B621v7bCsg17ReQBgBiTvBmgGaniQi98Hic8Ths8Tju8Hic27tZaCW3cfQjHGyPiWVq2YS+WbdiLhlBE9zePqmSL22+/HVu3bsXbb7/dilfddSFDhR2IbGuw4vE41r/8DPqNmQxvQf80ezTH5n31IlxDiWJqV1Lk9+KKcweL9gteRsgoK/bj88NBbK0OoLjQi+01yaLiedNHYWFC9kZBMnSWK6ssm/okMoScMJDL59qCgPDI1eek7eOVDcxCcFxBPNdzao3rMttHc2vZZk0cavrddSSsrqVkaF/UNISQ57CZhnzpGvfXn0jRH/R5nELGqS1ChQAwdepUnHHGGXj88cdx/vnnt8at6NKQocIOhFlPHrNQ4S9LYrjooosw8Hu/xvAx4wFY18ykCz/RvwBSQjfGkA6g1WEBeoHYVR/VwJ2o6XrhvS+EUaMHmgvw0r6obbn2WTvuKjtLTNjZ1CfxQlCaIM1qoIxjCodaCd0S5TudEK7VmMoJWpPEsGxDduK8mcatfV2Z7nW6cGlny3HR7413PaBogRk8TjtcDhsLkWvkGQoLAtoirj1EcB966CHMnz8f//nPfzBgwIBW339HQ4rsdgPDxfNEX7zwK/yrchv6XPM7XDqu0FIp3EodnibDIr8Xew5+icZIDGMLvTgSbEr5XMnQpDq8MWyy+LIxlkKhgGasFGjJ6nRUeRIWNd4Dfi3G60p376w+19wJN5dxR7Qeact9mOWfWuJxdTZQOxJuqLhsmRHG99pbTJqjvr4ehYWFuOeee7BgwYJ2OWZ7Qua4ugjMWIT0N4VY3v5oFyoqKuAtuRRxKM0KteyvP4GYqv3bFNXU3jlDkKN0WAGKC70AtIaRRf7keOnaXSIUMnvKcISjmnEKR+Mi/6UCutyXGYJN5jI5rQ2rfF+m/FQu49bYR2c6p54AI1sw3a/V63HotAmJAWomeGslytta7NaCggJcffXVeOKJJxCLZafg0V0hDVcHIhtixakH3kZ+fj7mz7lFFzYjA5fO+NEYSK6+eQLa6nNExz4SbNKNSerJ7bQlVtP0yKu6/BdX3DBDhBk2q2vJlMTm7y9duwvVDSEsXbtLN5ZoHWT6jlZvrcn6e+tIcOYgRzqOXkMoigLW6cCsmJ+Q7llsLdx+++3Yv38/1qxZ02r77IqQ5IxOjHg0gg0Vz+K6666DJ/8kAAcApJf/MZN1ShdG5EYzU+jt/tU7E2empPxLZA8AokeXFZz25FTBj7t5X33KdfXLz8P5D643TabTtq/v0O4LeYA0Nk4cxnGuCuhttY/Ock5L1+6yJC40Rx6quT3TWpI3TBeeXLZhrynFPd0ySwFQHwyLXJfV82L2mtk2LcX48eNx7rnn4vHHH0d5eXmr7berQRquDgSfWM3kmvZu2YhA/WHMnj0bt66ylnLKJOvUWnA5bEL12vh3XSCpkpHB4UIkpuJHz34gmHlcw814XaRAYTbZ0rbG86KxlTEmA1nkT+b5+Ptm27aHpl976gya5UmD4ailYWsOi7Gy6mi7G29uuCifBWgEin75eSICkS1UaLWKxlyWmbZge2H27Nm46aab8MUXX2Dw4MEdcg4dDWm4OhCZVmhzX/w1Ck8fiaKiIsw+flJWK0++j0yGMZv3+DjfZUdDKIJ8lx0rNlUhGNZyVZTfIjkoTwZyBmAuwJuN6Gy2kzCtvs3UHMhAHgk2pUxGVtu2h6Zfe+oMGun8syYOxY+e/cCypKErGW/Csg17dXJLdQFrozVjnB//2FqjW3Q5bQqgaDnbbEo70j1vrWnkvv3tb+PWW2/FqlWrcPvtt7fafrsSpOHqpIjHYjj6ySZc/N1Zzd5HJsOYa6iQ2pVXNzRi2Ya9Ipe1vSag6y80YkBvYcSsQKQPDrO6q+ZOwulgNJC5bNvZJ+1c9mFEc4WLO5vxJsyeMly04akPNsGmKIglSNQ8nO20KSgdViBqFqnDwYA+mnhudUMoK1JUe4QKAcDr9eKCCy5ARUWFNFwS7Y90K7Rjn3+M6Ilj+Dz/rJRtAeswCd+utR+YQp+Wx7Ir0IVdivxeXcdjPrbqlry1OiDYVnSuZsYz23xHLshF2aMti5jban/Z7sOIbBYvZuPWrmVr7v5KhvbF2YtfA6Cx/17fUWfp+XPPKhJXxXNDf1Md4Mbdh+BjxcSdBTNmzMC8efNw7NixLlP+05qQhqsDkW6F9uyjL2DXST7Mm1Vmum1HhArrEwzDmKo3TvvrT+gUNxysJxfvEMvmCgBJRYGGUARL1+4SWm65Ipc6o1kTh4p7UzK0rynxg48z3ffOMmnnug+zwuDOiFwEfs9/cL0IDfL2PNmARxQAiKJ8UskwevFmvzmqa2zrUCEAlJeX48c//jFee+01fPe7323VfXcFSMPVSfHBv15H3zO/Cput+RULxjCXFeMQyC1UCOiNE6A3UA67DZF45jqTBl0L9IjY/9K1u5DvcrQJy45PIBt3HxITjdU+OsuE3hyF9nTvUwFtZ2dW5jLhk2yT8beZCcWFXsyaOFRX9A8kQ43BcBQrNlXpmLncSNF5k/RTW4cKAWDYsGEYO3YsKioqpOGSaF9YeUQPvbgBtfs/Q//x1+TkEeXaSJI3uCP23OZ99ZZsRCvSxZCCXgjWHkMkrsJpUwTLD9AmhUz5LiPC0Tgmj+zbZjkfukcep11XRGq2j0z3vb3o8K29YqdFTGfrlpyuC7QVuDFRkb6O0GlXEImpQk+wIRQRtYqlwwpS2LikEm98voxGigxUpq7GrYkZM2bgiSeeQDQahcPRs6bynnW1nQxWocKhNXuwLc+F08dNzMkjyrSdMYxoZFzFVI3tZ0ZjXrp2l67/Fp8cSFtv9dYaXDz2FJQOKxAG8YpzB4vWKNnC5bBZirO2Rs6Hrj8YjooGmlb7aO1QIafXm7Eh+f6MVPzWRFve39bcn1lIznjPqaiY1wdaIvGzVRQgEIrA47SLkLFZqI8K9o3PEJ1DR2LGjBm4//778c477+BrX/tah55Le0Marg6EWfhn1sShuOCCxSg6dxLseZ42O3bJ0L4iGa09hCqqGxoxyOs2nazpoQY0kgUPx5QV+3Vt6EuHFehWqrl2Qs532dvUI6D7nA39O9sQXab3zej1mdiQfNtsJvBc8mFtdX+tShP4OJfaODOP1+iVUlExV2SxAv1mad0VisREyNgs1GelwMLDzPRs0LZm3ldbYPz48Rg0aBAqKiqk4ZLoWBw5cgRvvfUWhl32ExzNkTyRKwGDUFl1FFRAzAuJOahtSUzVqMQOu2a4FGghFlLKBpCyUr1/9Y6MdV0c1Q2NqAtox6K26Bt3H0K+y9EqHgG/7uaw71qC5lLxs5nAcw1NtoXHlc4YN6c2Lhum6cJXtkFFMgzYHBCL0PidZxMq5MatLQkZRthsNlx66aWoqKjAb3/72zY9VmeD1CrsQJiJck6cvRTxeBwXTJ2uy7+UDE1tYkist+Y0kpw9ZTh8Hqeg+pYV+2FXNO+JP3w03rj7EHq7nVBAK1wtLKNCe1ipCHlIQS+UDO0LBUCeQxMnHTGgNwCtXqbQ587q3sRUPX0+HI1bXlumMRed5chmH62N5lLxM11Xa9yb1r6/Vmjp/j4/HERdIIRlGz7F3QmjBaR6XIphPGOcX/z+PE4bFGhtSkg4lxpCkmG3ChUW+jyYN30U/j3/QsyaOBTzpo8Sr7Xlb8cMM2bMwO7du7FrV8/S5pQeVwfCjJxRs/VfOOnUUfjkmAMxNWJJrGgN5LscupAG7dtKDojCKfkuh8gPkTGlVigfVwewvSaQkMqJJUKFWkgyGldzChvyaSgUiWHzvvoWX7Mx5Jbt9q1Fe+c5klyp+D0RZp4mhbgz5U1VJI2XClosaK9QBKAxEhORBAVa2xL6bmjRyJ+RjpR6MsM3vvENeDweVFRU4Kc//WlHn067Qfbj6kAY+3H9aeNu3DLtbHzrhtsx88Yf6t6zajIJZNdIcvFlY8Tn7ArQ2+0UNSrGz5mFZ3hPr/31J4RSvCvhVfH8F/9BFRd6sZMxDoHM6vGZwBlhfGLqrD2u+JgkprJpoMnlqNI1CO3o6+L7yIZEZPX7NdtfyVCtGSjvJacomfUwCR6nXTBcqTEqLbIACJUMQjb9yYyw+i7bq0/XZZddhvr6evzrX/9ql+O1FXKZw6XH1Ymwe2sl4k0hnHN++h98tsxBIMm4otd4WwermhOzVW5NQwgqNI+KHnMF2sqV9ysyzieaHJQdkXgMDrsWmY7EY/AkOiEvrtiOSFzVGbhMUJE0lHzrziCjlK2WXiYpKWMOzGz7znJdmcgUxnGma+f7o0gDiS0DQDZLbVpANUZiYrx6aw16u5267fJdDuS7HGJhlo00lhHpnsf2QHl5OX7wgx8gEAjA6/W223E7EtJwdSCMocK/rl4PJc+D1f9xQqmxVr2g2DqBPDIKPfFwFD2o9CDR35NH9sfG3YdM67jMJhYV2oShKMmJQ1EoSZqcSfgKVoFeDqoxEoPDRqEabRUcTWzLDWJzUVl1FMFwVLAbiczR2SjfdK65EEPMtu8s18X3kY1BzkVGi7MfKTRo9OqN4O/z7WKq1sRUQVKrcEhBLxwJNolmq80h56RjnrYHzjvvPMTjcWzZsgVf//rX2/XYHQVpuDoQxpXaoKZaNPjPwO0XjgSQ3pMyIh3rjMpbqHbLriTpvPQ6r+Ma5PWkPMQiJMdmAlUF4tDChRQynDyyv5hwKquOYn/9Cd0kwj2qZRv2CkPIDWLzoYprCjZFhXp9b7cmR6WFcNwZi47Nxq3lcVHIt19+Xtbbcvp6a50T5TFbI3fH95GJMZhrbRztj+ezMv1M0r0fjalQkaTQkydnVVDMYQztG//OBdn0Kst2n6NGjUJ+fj4qKyt7jOGSOa5OhIGnDoX9tHPxm6X/DQA5PSRUk0RyNZQXoAeTHkoKFfo8yZDJvOmjdHI3vICYPLNgOIpoXMXYwmSuYWyh1s+K8hC0XyJvNCQKPMm7MuK+mWMElTnTKrq14fM4BdWZj9s6x9Wa+2vJPtoq/9KaE/KKTVUZSykytdCxKdrCKhyNIa7S3zZE4yqiMVX8ho3nlU6L0CwvnOv9zDXHmQmTJk3C4MGD8eyzz+Z0Hp0JbZ7jeuyxx/Db3/4WdXV1GDduHH7/+9+jtLTUdNspU6bgzTffTHn9kksuwerVqwEABw4cwM9//nO89tpraGhowOTJk/H73/8eI0aMaM7pdUkcPXoUB6v3o985l2dVg2X2UFGIjCbh/fUnMoYKrRAIRaACeH1HHRojcWFU9tefQFM0BjUxBvSCpvRwO21KwtPThxGBpNe1eV89xiYkocYWerG/PqjTPGwrKOz67IpGtSevMhsvwOz9lngpHbWPzo6la3elNUo+j0PX9RpIXQDFVS1ErbK/m6JxLe8ai2F//Ql8uOiilH2bRTBaU4sw1xxnxv2VlGDNmjU5n0dXRc6G6/nnn8fcuXPxxBNPYMKECXj44Ycxffp07Nq1CwMGDEjZ/qWXXkJTU5P4+8iRIxg3bhwuv/xyAICqqpg5cyacTidWrlyJPn364KGHHsLUqVOxY8cO5Ofnt+Dyug62bNkCABg8oigj6SLdQ0UGjMKDXN2aq11XVh1NK/lED7rZxGHcPwcZJfo3ykKDRuIFT5bvOXg8pyLlloCfRR7TVZRoObIhZ5gprfNOxYBmtIy/LSPMFjlmXrvxtVjCmKWDWTizNWWeWrv4vaSkBI888kiPIWjkHCqcMGECzj33XDz66KMAgHg8jsGDB+OHP/wh5s+fn/HzDz/8MO655x7U1tYiPz8fu3fvxqhRo7Bt2zYUFRWJfQ4aNAgPPPAAbrnllqzOq6uHCn/zm9/g3nvvRUNDA+x2e9ptzcKCZcV+PHL1OVm/x8OBZqFCUmrXNdyzK1hUXqR7oBeyAlAO/jkrUH+vzoSeECqkcG5beHhGhXWz8DWF5uhzXE6MQtiZjJYZzH5zRBZSAJSP84sow5CCXkJj09iJG8guVHj24tdEbeO86aNyCoW2ZkgVALZv344xY8Zg/fr1XTbP1WahwqamJlRWVmLBggXiNZvNhqlTp+Kdd97Jah/Lly/HVVddJTypcDgMAHC7k4oKNpsNLpcLb731lqXhCofD4rOAdtFdGe+//z7OOeecjEYLSL9ay/a90mEFlgXIHLqGeyZyOm6LHJaV0eKhnGyLkTPlMVoXqjCm/fLzsLNW+131djtwvDHaKnT4ukDIdAJvr1BhMBzNWR4qF4mpdJ4E5VyN0QL6XQRCEbidzRP0MVuCO+wK4qqKsmK/iDIU+jw4EmxK69lkEyq02j4bg8OZiKSc0xKceeaZ6NWrV48haORkuA4fPoxYLIaBAwfqXh84cCA++eSTjJ/fvHkztm3bhuXLl4vXzjzzTAwZMgQLFizAk08+ifz8fPzud7/Df/7zH9TW1lrua8mSJVi8eHEup9+pUVlZiRkzZmS1bS51XNmEG9PpGBph3C7XMBsX53XYFcTiKuKq3jhl460RWtuocQ+Qt2OhsBTPIfJxrhTyI8EmXfI9115bzRH4pXE24sJm45awGD8/HER1QwhxNSkWRvd6SEEvUcOnQstBZQv6rRQXauGxrdUBQ1Gxkpaq3y8/z1TBxOo54veSe1l8m1xhFV7NxeOy2+04++yzUVlZmfPxuyLalQ6/fPlyjB07VkfkcDqdeOmll3DzzTejoKAAdrsdU6dOxcUXX4x0UcwFCxZg7ty54u9jx45h8ODBbXr+bYWjR4/is88+Q0lJSbsd0/hgGv82DRXalJTtrEKFVuB5LoctKYrKjY/RaDVGrBPwBfku1Aeb2iVPpUAfwgo2RQWtuiXFyO2NbPMrrVnzVRfQZJroe7NaINBnsoECoE9CAYaIQgBEaFCFlssiPU4zqr6ZB7hsg3mtZFuhtQqYx48f32MIGjkZrpNPPhl2ux0HDhzQvX7gwAEMGjQo7WeDwSCee+45/OpXv0p5r6SkBB9++CECgQCamprQv39/TJgwAePHj7fcn8vlgsvlyuX0Oy2ImJGt4Uq3Qsv2vUwPBj34LodNsArzXak/FwoVepx2FOQ7xYRUnKDNhyIxYfz4v8Y1CTdGRmaY3+dGfbAJjZF4gn2oSU65HHbMnqKpzxOMEj7c8LbUOzPOpzx0uiohI6RCY2LScdpTeT5bZKqjagsW4+Z99Vj1UY2o1/PnkN9MlWXSPuv3uUW4ORyN6wgXXlbiEEi0HBlS0Et4WXsOHhcF8vvrTyAYjmLyyP4IhrX6vx89+4HuWo3kEZ7zaomnBLReAXNPImjkZLjy8vJQUlKCdevWYebMmQA0IsW6deswZ86ctJ998cUXEQ6HMWvWLMtt6Gbv2bMH77//Pu69995cTq/LorKyEieddBJGjhyZ1fZtHSqkTrJaSEcRYwqNmIUKCxIFtYDGODzCvCAyUjT3qCoSIrxJI+JgLSkcdgXxuIo8h2YUuWICKWx4nDYcb4zghfe+0O2HsxgV6A2k1pk5NRyZYOlnHZ40Aze63EM0K/xdunYX6oNNOHvxa7qyBD4ptjV1vqWgydZIMuAtPoxdtTfuPiTKEOqD2ZMv+HfqcdpAv0z+uyDKuwLNaNG9pLrFhlAExxPeFZeP2l9/wjT0S3qGS9fuEq8Bqbm5XIqW+TYtLWA2Ay18e4KCRs6swueffx7XX389nnzySZSWluLhhx/GCy+8gE8++QQDBw7Eddddh8LCQixZskT3ua997WsoLCzEc889l7LPF198Ef3798eQIUPw8ccf48c//jFKSkrw97//Pevz6sqswiuvvBK1tbXYuHFjq+43XaU/Z38ZGYe8sR4AMf5w0UWWE1U6ZqLTpiXI6eE3ykKpMPe4eGFwa4P3bkpXJJ0N+OeddgUDervEpEolCpx9xvUi+XV3Rgai1T7SiQXTtRCZgySWzK6Z4PM4sqrhc9oURBO5MAA6wWhqy8MN9P2rd4qIwIgBJwmhaCqi5+Qis9+B8TUS6uXPTjqkE+BtSQGzGWKxGPr06YPFixdj3rx5Ld5feyOXOTxn+s6VV16JpUuX4p577sHZZ5+NDz/8EK+++qogbOzfvz+FVLFr1y689dZbuPnmm033WVtbi2uvvRZnnnkmfvSjH+Haa6/t0hXguaKysjLn/JZZLy/+N5AaUuR/847FKzZVCeWLyqqjmDyyP+yKVqTMx81FTFVhI7fGAK/HofuXxnZFS9pTT6/WRlRnDVWYn122SO4rGtO3bjGuwCnPAiQ7SRNKhvbV9V2j8SCvO+W1XMetvQ9+fbyflifBCKxu0MgcZKT47TZbhwSyMFoKIIyWAm1h09vtxOSR/cWY99RatmGvMDpN0ZjQI+RF9C6HTdQk8rHV2VZWHc2pp5rx/hgLmFuzd1dPImhIyacORiwWg8vlwqOPPorbbrst688ZV2tmqzdjYSf3iLhXle9yiBX14svGiOJPM4+L167Qe4C28o2rycfcl2g7YgazFbcZaIXfFuA1ZOnONVcYz5lW6Ks+qoHbace00QN1OTDufRq9MwCidYvRA00nWdUe+yAMKeiFj6sDcDttmDZ6kK5lSFuAPDMj85T/psi7+rg6oPP0PE47mqIxEYKm3zugD/ORR8y9OWMIsjkhvtbUOrTCTTfdhO3bt+Pdd99tlf21J2Rbky4EKjHw+/05fc5I7TX+DcCUrEErxSEFvdBQHcCQgl447eR81DSEhHdD4rT0r9lr/D3arxnpAkhlJl489hQxwRUmEuzkeZAeImkgvr7jgGkYzyzEmCncRNs6bYrOK5o8sn+rTbhGQ1vxUY04bigS08ljAUY1kaRIcDgaF9ftsCtw2BSdceULBjouderleSaC22lHQX5eirKK2Zjvg58HfY+680gwAkOReMZ7mO2CJR3o+zXmI/mfoUhM5ELp+7ArlOOMweWwoSA/T2cwjIYjExW+OYYmXV65tQyX3+/HG2+80Sr76sxoXqWfRKuhpkZ72E855ZScPmcMV1iFL6xCFdtrtAlne01ASDxRx2Iu22SUcDJ7D9BW6dxozZs+SqzMXQ6bCMXluxwoHVYgJsv6YJNQjc93OeD1OLG//oSYNKyMFhk6PmFlCjfRtpG4qgvRvb6jLu3nWgrVYmwEZ9lFY4xwElNTSgIIxlAZAF14l9rK31V2Fv49/0Lx9+LLxqSMzfbBzzimtszwZPqs0yKczME3cbK4nsdpR6EvKWLgYO+RZ0XXO2/6KEFQWrGpKiXMPmviUEGH5+OWoq1DhYA2j9TW1iIeb6+C/Y6BNFwdDMoH5mq4eA5ixaYqBMNRUa+SDcqK/SKZPXvKcPg8TvF5ylN4nDbdGEhOHDYlOaYpwp+YOJw2BZv31SMcjemYiYAWWqI8j11JrdFqCEUExXj1VvMVvNfj1DHNCBrb0J4xX6UAoqkloGc3Gj/Lw2JGUNFrW4AvClSkTsSEhlBEsOHovvH8ZS4w2wegZGVQWgORuKozTGZHtSuKeC8/LxkwcjlsCIZj4j2+MDE2jwT03k+2xfcthdEItqZRJPj9fkSjURw5cqTV9tkZIUOFHYyamhooipKiRpIJ3MPiUjZmIQ+zuq7ZU4YLiSf6DG0zbfQgrN5ag2mjtdo8Pqb5NKmAoE2slCMDtAmIyB7JLTR8XB3AvTPHiFAUz5FwajgxH3lIkTySySP74/PDQVG4yuuzXA4bXA4bwtE4orE4InFVW5mrEOP8PIcu1MlDmeXj/Kisqkd1QyN8nvSPx56DX6Z9vzXBi7V5OFKBPnRXHwyLe9EQighm59K1u7B07S5BRye2Hb1HrwNJZiCQuzJKS8HXI0aWKaCFknke0Sz86HbadAuihkQdFx9PHtlf1HTtrz8Bn8eJfvl5GL5gdYrOIvUusyovaO/GkelAC+Camhr07998QlVnhzRcHYza2lr0798fTqf1yt4M/fLzUN2gPXgAdGMOq7quTHVcfMUeUzVyAYX4yFsq8mstSbQQjYKSoX11IqZJRYTk6tfttGPzvvrkZJtQPAhF4lj1UQ3unalPlhOJJBiOCRIFdTcG9N4HDVMU69lMH4mpKe/zyZLauGj7SR96bK9J3a6Yq/QDqZO21XbGa+bn3lrElLaC3+fGwWNhROIq/D638Kw44WjyyP7CwATDUYQicdgVCCKGEeRRUi7MriTH5OnzbgnpdBs7k+GiXHltbS3GjRvXwWfTdpCGq4NRU1OTMzED0B4y/q9xDKSymAB9UtmqONksua9CW5mTDaBCTg1KSh3K2YtfS7yTTIoDwLTRA1NCgCr7lxd8LtuwV/RbCkfjcDmS4T0jgYAweWR/g7dnDqMaAyEUiWdd18VrwbLdf3PQVszKtkY2SiXZ6FLy3B+vj6MFF0UeHrn6HNOiaLPx/at3iuPHVO0/+j6L/FoIeGu1VvN12sn5KZJd1FW7X36ergbykavPaRO2YLYgBaN0Oq/dATLH1cGora3NOb8FJGP4DpuiG3NkG7s3Pmg8iU1Je2Ouh9OMzXT3yJi4nXZdDVhl1VEM8mq5sEKfG0MKegHQDJyCZFNH2i8ZK5fDpjsvmhA0byGpTbdx9yEsvmwMCn0ecc52JTVXxeWreA7H47ThrrKzxDXzPJaHqZZ7nHbd/ea3vtDnNs3BtRTtlGpqNWQyWgqSOatcr42zaY21ZdmAflf0NdmV5POzv/6E0D78uDogvLPVW2vw+o66hBJHVLxf8VGNeB/I/rlrC+Tl5eHkk08WpK/uCmm4Ohg1NTU5G64Vm6qYwrotGSNT9IXI6R7qdMlpI6vqw0UX4cNFF+kMEJ+XzQoxucGh94lqTVT0moZG4bWpSGojUvivsupoirHiyexwNOkVUSuMcDSGWROHihW2x2lDXqJ1u8dpg9OuQIHmmZFROuuUPmI8YkBvYbzzXQ7sOXhcHIMmYo/TppOPMt4PYkpmg9SC13TbdjHLlQHEJgWS90+Btiig8YxxfnGPfB6n7rdB/xrZtFa/bT6m31X5OL9gVrpMit3pe6QQOYWRnXatw7fbmfxMnsOW8tyZCQO0NYhZ2J0hDVcHo7a2NudQoX4lp4pwVSSmYunaXahuCGHp2l1p2WXpKvqtwCcH8lIUaA8oPZz0oA4p6JXwoGLol58HBdqDDeiZieR9ARAU7t5uZ1YMSZpoXA67brxiUxUWrdyGhlAETYk6pFAkjsZIHJGYprxQWXVUVxJA44+rAzj/wfXiPpp5DY2ReNq8UGMknrUSRy5hQKMX15Zsv1z23ZqnoRErYmK8cfch5Dls8Di1xQeNSQiX2LT8N2hU/VCgkVbod1gfDGPzvnoAwOeHg0JP0agYoxlRm+43SczZAb1dGOT1YMSAk8S5hyJxLF27C6sSHtjG3Yc6xPuShkuiTRGPx1FXV5ezx2Wc0GmScdoUBJsSBcJNUV1+yLjy4wKofJzpuBRC4/Rso8xOdUMI22sCQkiXj5dt2KtjJtYFkvmLxZeNEQzJfJcDsyYO1Rlifg0rNlUJuv2Qgl7C+wpHY7pcHDeMSiJk6HHaEAxHUeT3CoNK2ymKlnw3FliLfUBP+yfmoYLkBO71OFpcaGt2XIchpGmVQ+MhTV7bBOgNkjG8yT9nVNhPh9aOigoGYYJIoy06tMUHjRtCEaz6qEYsIFZvrRG/QSPj1ux3SNtvrQ7oSBjGzzVF44IgdLwxIqIF1Q2N4nfOEU5ISWnjeErUoz08ML/fL0OFEm2Hw4cPIxqNNoucQYjyQmAFOu+LilijsXjKyo8IDKu31ujGRuNgNHZUZ0XN/lRoD6hxhWtTFFEgSmPyzsYmwnJjC70iEV5c6MWsiUNTwiwB5tkYwz6hRLuV7TUB4RkZPaS6QGOStq8CHy66CAX5LjSEIjgSbILf50EoEhcGlIqouXfDc3xej1NsG4mrouhZRXICz0YsNlfwsBqQPn/kciRr2YLhmAi9AfrcHjc4NQ2NouTBiDZI1+k8Un5+/B1j7zbj2KZoXroxL2rmcXmcdrFQobGW+9TuR5Hfm/I5QPOIaQEYU5P1dL6EnmaR36srhOb3nsLkPJTZHh5YT/C4JKuwA9Fc1Qz+o+disVFD3Ikm32hcTZGEIiq7kUFFOoXGupdZE4fq2ID8UMQapIfTOMnyMQ83Hgk2pYzNVr1AUuTX2FGXq7HThDGkoBeCiY66RX4vjgTDon8TYF4iQHU51JnX73Pj4PEwIjFVEEhIIov2B+g7OhvBa89sCmC3WbMQWxP5LrvwRBpCEUPYz/z4KoB/sO+X10eZ1Uq1FHx/9PvhskwalV3LIxo1M2lM2pguhw1N0ZhYfPV2O3VjFVrbE6K7N0Vj2HPweArJgl7TdCWTa3quYkL1dMcbo6KT9YDeblHHSFqG9PyUDO0renxRGLOtG4n6/X7U1tZCVVUo3SwvSpAeVweCGnLmWnzMQ3bl4/zwODVJJS3RnGTs0UTt97lTVn5kKI4Em3RjHnKjcTAcxfkPrhfhNB4iK/S5RaLbqBBuhn75eagPhkX79pqGEDxOu8hR8DwEr0szqnJTiCYeV5PeUuK/7TUBYUyOBJvE+zwsaUTpsALMnjJcGJqahkZhZHgObHtNQLcfRxp2RTAc062+jQuLtoKxQSM3rOm8QbPiX+O4pSAbSr8Rp11BvkvzuIr8XrFIKPJ7MWJAbwDaQoReH1LQS+SixhZ6UejziByUglQPCdC8Mvpt0OtGxRbNuMVTxoDe+3Y57Cj0eVBW7E/JES++bIyuKSSvOSRFklyU5ZuLQYMGoampCQ0NDW12jI6G9Lg6EOFwGADg8Xhy+hx/OFZsqtJV8A+bvxoAdCKyNQ2NuHfmmKwaTi5duwuhSFyQHUKROKJxNUUZ43gjTYBKSgdXCoeYgTfxowk2FImJPEJdIKTLSQBJNiI/zxfe+yLFY6RC6H75eaL4uV9+Hg4ea0RM1byvHz37gVDj4OoRi1Zu00kDcU/K6JXSmLo8WyEcjesUH7gBaEvl+84EYy0X2QB6LRJTxe+A54vSjY8Em4S3Q50RGkIRof7OTS0VJt+/eofooE0CzpVVR1OK5jUVl6S6vHHMSzE40tVrpWvu2hZwu7UFJs0v3RHS4+pARKPa5J+ragYHj5mv2FQlHlmHLakP6Hba8MJ7X6C6IYQX3vsCP3r2Ayx8ZRvqWaiOYNaLi1a2ZcV+4enxFSdgTsMvLvSK7WckaMdlxX7hhTgN3orP4xSsMZ/HKfIQZcX+FMFTM4+RCqGpBgfQe1/760/owp281TtfnQNaLojGRq+U9r+//oSuKDoV1mHcnmC0FEBHMeeeOGelUpSgyJ/MeRb5vTo9TRoX+b0pTMKSoX1R6PPA5bAhpmrH5LT5WROHoiDfBRVJY1c6rACA5mlT6cP++hMilEjjgvw8HEmUNxTk51kaKGM+uDX62DUXDofmj9D80h0hDVcHgn5Y9EPLFla1WlYJ3xEDegsPZGt1AKu31ujU4Mn4LV27S9dQkuj0mTT5iH5u3Mf++hPIdzkwb/ooMVEAmiFVoFGKKVk+eWR/BBIMMsLHCcZXZdXRlInBqukhoFfSoInQ+HdxoVeQRAp9bl3Iyedx6saUpyCKPmdr8onJriQnZ2MAcaxBkJeMttOu6IqjFegNum7cwRXIRhai1Xs2RftO3U5bon7OzsSa7bp7VD7OLySceOHv/voTKB1WgEFeD0qHFYjx/voTgnVKUky847HP48SQgl6Iq9qxrJTYzeq7AL0nnE25CP0uif26aOU2rNhUpStFaW9BX1oId2fDJUOFHYjmGi6rHluzpwwXoqo8Lr+9JiBCNlQ82xCKQoE+BEcahQASvbC0dh+hSAzVDSFd+I/CbUvX7kI4kRgHkDKm2D6g13jT/tZCRE3RmI6IoVGftcmMwoRGbUVjroCPKcGvADrva/LI/uKaeGHxwWNhDOjjFuFKuwLLsQZV/Mtr5KjFCL1GBs54HlqI8TgiMa29yrzpo7DwlW1QobEWw9EYIjFVTPDtQejIBkYWIgdvaRNXkfiNZdZA5NqQ4WhMl1syE4gm423U6OTNT4n23hiJCXHhySP7IxiO4v7VO3H/6h2IxlSWZ22Ex2kXoUJaSFVWHcXmffU6w0hSZvQ9c3FiQPsdLlq5TXT8DkfjmDe9fUOFNJ9EIp1bg7IlkIarA9HcUKFVzJzXYfkZo62s2C8MDdXCAMnpl3+ettu4+5CYRCh3YDZ/hqMxnZdkDIkZV7n98vOEGnxBfh6qGxqFHhwZRi9rCGmzKbqJx6zwdOPuQ2J89uLXRH2XCiDAOvdWVh01pc1HEjk8UmbgYUPjeNmGvXA57Lo8IJBUaacWIzE16X2p0Kt8ELuN7hcZLUAjwlBo00h5z0X7MBsNwFzBtRlVJGva6gLad7iz7pjOyGajVdiYKGnQGIXafaX1AZd1ou+YyiOMeS+enywr9mP11hrkMeNppV/JF09k8GghZRTTBZKCxOmaZsZUIBZLLm6MOeC21i/sCaFCabg6ELQisttTpWbSwepBWLRS87bsSpJBZ1eAR64+R7TqoBAPjY2q8IC+XQmgCeMCyYdVL0Kr6CZJ/eSaKr67aOW2RL7AJY63v/6EjhrPmW+k5s4NAk1CVp4RkNqWPt/lSIir7hCeJ6AZB6dNwYA+bktRVp7AJwNPiuSAdj5Ffi/2158QK3AFSWYan5QBLVRKYyNpo7WEeeOq3nBYif5ylX/jdkZKvMNA5/f7NFKRFS2cflt2RfNG6Xvi47Jiv86jod9hQyiCyqqjOvJFoc8jug6YEWaoXKF0WAEqq46yRVKSYOFgLW48TjsK8p1i8cT3R8K6g7yaYSYjWtMQyoplmVz0tH94tycYLpnj6kDEYjEoigKbrXW+Bp7Y5mMAIo8QDMd0Y54fCrOiYqIfA5pBoIlbEVtocDlsujART8AbiQuk5A3Q8ZLUe17XYkzlKNCUQOizvd3ag2mkPoejcZEzmzd9lBDbnTd9lDCgLodG/Jg2epAYL5pRZNrxlsZGNQ9jrRlN3Pkuh/AevB6nGPPuu/fNHIO7ykaLhpd+n1vLa9mSBdtWY97llzf29DjtKYoY/BbaFejawHDx4cWXjRGLBuN2984coyu8Jg/T47SZ5n+MtHD6m/pbEamHxosvG6PLfRImj+xv2SnYjJSzv/6EMLwfVwdEvpV7UES2iMZUEUZvisYEQYnquGhMOaqahkaRq6JwtgKdPKjpmNYI6ck7bQNaCEvDJdEmcDgcUFW11dpsW9VmpQNPIvNCS94iJRyNCykph13R5c/mTR8l6sUKfW5Re+O0K6IYkzB7ynAd24vrC/K6Fv6we5w2MeEQuEdG21LxKk1UZli2Ya+Q7uEyPgDSKoaka7luNXGToaJJ3IimhDRQXaAxobqhio6+0bgqrpeP+URLnlQftxN3lZ2FeGL14PM4hadEhplybzQeUtBLeDsARNjVaGCWbdgrjAh5l2T0zWDV4Zcbfj4GoCP10Dhd52az+83hdtqEN05MxGzqubScmD6E7kssPoDEMxCOmhZmexMhZhpz8V7jM9AeiMW033+uufOuBEVV+Xq56+LYsWPwer0IBALo06dPR59OVvjLX/6C66+/HuFwGHl5qU0gc0W6PkRUyc/JA/Omj8KildvEg85zADz8ly5XMWNcMn/GH2oAuM9QO7Z07S4Ew1FE4yrKx/lFF+PiQi+uOHewOBezMJ+x3mZ7TcC0Oy3tgya06oYQCn0e/Hv+hYL9aDxGoc8jwlNGlYYPF13U4vs/a+JQnP/genEudF40uVLok5+TFYz3WCPbaEoZCpJhSAUak5E8EWP4lz7blJBL4u/zMd3/+mATQhGNIdgY0YwuH/s8ThEetAqz8t8e75tmdm4AdN8FEsejhYpxX8ax5s1rd8vlsBsam2q/b1XV51N9bFxc6NXVHFII1XifvB5nyvE7ujPyv//9b0yaNAk7duzAWWed1SHn0BzkMod3X5PcBcBj0a1huKxAjCsFGouLJ8DpwTROlnw5k07xgddFGeWPjKwwzjKrrDqKuoBGxtAS7YNFgt1oXKk5oD4X0oTSYQUp+T46rhWDK8+htXXPd9kRSBStlgztK4wvJ0eQYki63Bc/JzKKvBkmsdqIFh4MRzGkoBfqAiHReNBofM0WGTQh0jHrg2GxmCDCAm9Z702w6whm3yARI4zv05i84+qGkPi9kKEC9F2Ug+Go6fUTucEsH0mv85wXyTgR+G+Gs0052YLnP/mYUJDvShHDNdOV5GNutADN86XFHV9gcC+y0OfRdUvuKMPVXLZyV0L3vbIugNZOoprRhznpQkUyxKQmtqdVJkk40XhIQb7whvhK1Qie2Nd6g2mtQ5x2JUWXjdPpjYoWFMYzp59bX5txcrAirnBvS/u8Rl5pZBqLgL6MgBRDjIaIxpSoJ0PDvQdeEhBiJI1QJIbjNcn6tHTnTOe9bMNelA4rEN11K6uOYsSA3onJWGETfXI8eWR/4dH6PA4EQlFBUqDvy+9zo6ahUROtVTQijJON8xKeyvGagOj4y9mqnGXIFyxc9JY+F1Mhws08H0nHCFQHkOdIkiWKCzWCxKqPaoT3PzahVNIYiaHInxzz4/ExoT7YJM6Dvgfahn7/GkFH07x02hScdUofES5XE/eqPhhBYySGIQW9sLPuGGIxFf3y83DayfmoaQjh4PFGIPEb6JefJxY97W3AiPTVEmGDzg6Z4+pAtHa9hVXuxYqpNnvKcCHddLwxqhvzHFm6wtf99Sd0eny8GzMRF1ZvrUnJW3xcrdcTpPyXGf18xaYqXQFwc7recmIIT6I7bIpuP35WWEvEiYZQcgXPx3RbwtG4ZcdnTpMY5HXDriT/zeb8jQabt43RzsPMV9KMIn2HxxujOhUIAuXXIiyPxsehSEwchyb9GiZm7LApYuxxJhuA8mag3IvhnjvlI+kYNKb6sP31JwQZIpbwwvbXnxC5QT5ujMR0HqERoUjMUqORfvMD+mhyYTQmtQw647pAI0IJb3N7TUCnY0nnGYlpnRpUaL9vs3Y87QHpcUm0KVq7wt1q5X7Pym2Iq8mJlsazJg411fzjdOCSoX11quEEYq4Fm6Li4dbnETSDUN0Q0rH/CIqSDEeWDO2LzfvqhdoBLwQtGdpXl4cD9IQSY17JKsxGhg+ACJsCyVAPgQvo0sRuheQ8rDceegHV5Ht1gUbBVFMT1/GjZz/A6q01ghhhHA/yuoXoMNWveZx2RGPxBCGDLyqSYzKKZpRu8nwHeZPeE/8++JhvQ+/FVT3F364A00YPMs058Xwkz2kCGn2dq/s7GVVdUyXpq/PSyZui3wHdf8qxkSdMnqBd0VrqROOqyGVxb8rttGPa6IEpdHzKy2nMTS1PFo2piKlaUThXoqE8qxYOjoNyatzTNhbPtzWk4ZJoU7hcLgBAKGQuSNsa+NGzH+gaNxJUVQtDUex/f/0JwbAjViKFs8wcNnotxlbRnInoctgsKfR0fALlu1SYF4LyPNyyDXt1NHqrPBrPgfC6ryK/Vxf6Ito/TZA8pDTI6xZGxgzmtTpJg01/E8z2x3uh0TXycVKIOK67Dro23tKDj3muhxtK3o+LG2n+HadTyIgncjxknDgpg+d5aHzFuYNFPpJ+D/y3xtX99blUVZc/BTRvio79+o46XQH0kIJe2HPwOBojcVFvNsibJMJwz4prDz5y9Tm6YxCRhs6f6sjo+yzId+nKByjPCiQXTfOmj9LV+hlzmG0Nmk9IbLc7QoYKOxCDBmm04rq6ujY7hvHhJ/gTxcfGFSxJLPFQI3lqfHq2KRobj2vp8RYf6WjAHqcNYwupqZ8Ns6cMt6xB4+E0Cq8RAcHlsKeER6nFS5HfqxvTZG8MffGWJYA+tEUeFw8t8jGv1SG2GzUPpPPl1H7uwVGRrpmQLB9zoWISNuZix8Z6NRrz79PNGjXyljh8PzwczMe8LxUh3+XQFXsD0CmZ8LCulTaf2fdNwtDkzdH5k4dG92711hrRRJS8WGomqoUO4ynnYtaGxMqQGEPRyWaUNvF5HsoGkqUWJHHGowIc7RE2rKurg8vlgtfrbbNjdDSk4epAUAPJtmyzbRSZJdQFGoXem01J9kSyJRrPbd5Xj7pACJv31cPOOs4m81na9pTTUgxjAGKfmveUDBU2ReNMiTuOzfvqhbe25+CXYry//oTuwc9z2PD6jjodAcGIpkQrES0HkhznOZJGjCZqBZqRoHvktCuiwNdpV4TxUJFaq+PzOOFOqNjzGqd500eJia+s2I/JI/undOB12hWdRFFrIxiOChZjnsOGaEwjQgwp6KXzCgDgeGNEO//El+a0K7ox1eURFKQWj9OkbVanxQ0aSXFNHtlfJ6YrdBwTh85z2DCkoJcYc5HnsEF6rMivdS6gewto35FZrV62MOpgUg6L/4ZJgJnX9tEiwGgQ21tkt6amBqecckq3bSIJyDquDkU8Hofb7cbvfvc73HHHHW1yjLMXvyZWxlzOZ8Y4P1Z9VGMaBvN5nOIzWoRfD05htkKhz2PZk4uOQaEsq95UvEaJwM/Hqv5Kgb6eycuux6y+y6zGilbkVvVvNAFluz+ze2Z1n83uudXrVnJKZjDeO05VJxo6n/wVaIsV43fDz9tpV+CwKUJGa8SA3kKLkr4743nZlWR3Y0+Cwk/H4x5puno2fv28Fo7LkfEx7Zv/TjTJpzxTzU/KNVL+iqS8tHPX9mtVu2bcBwDd/nh9X1vghhtuwO7du/H222+3yf7bCrKOq4vAZrNh0KBBqK2tbfNjGQtrH7n6HPxjaw3Mli2cSMET9VSwSg9jOqFRLcyipugj0j6MOn9CK8/O9fAUDCnohYbqgJhQOR27X36eMI7haBzRhAKJFrLUPkEsP60duz1lgjEyFsnTpDYxVhR9I9Wfj83KAMwIKvxvXgPHx7wo1lgnB5jLXlmBGyGtHkz7fnq7HQiGY4jEVfh9bhw8FtaNiZQQisRTeqhFY6r4vqj5p5oYU1+sJPVcE8nV8nF2oVdJCww3I0NoUHR1Y2aIqcnfoXZqauKT+rHbpIi+MdH1wFhiQfslSS++EIipyfo1sxyr2T74+JGrz2lzgkZNTQ38fn+bHqOjIUOFHYxTTjmlTUOFZo0hJ4/sjxWbqizVw3lehhu2UCSue6DTobLqqKk+YqNhHzFVT6nnFGuXwybyT2riP04W4Lkpl8OWlEmKqSk5JyJ7GMHDXEYdQjI8RuNgbMFuFloy3idjvku7F0kad5SpMvBxIJQUPqYcIuVbaMwn6HTaePz71ggx2r1sCCWLrusCjaZjmvTjcVUQWpx2RZQP0PnaFE1bkUKTeQ67GOe7HDoppH/Pv1BIY3k9ToQicRTk56Eg3yXG984cIxYLvG+ZGfIcdvEb0PJ6ycXLtNGDUp4Dvy9ZlmDV3y1dyUI0ropwe7/8vLT7yLV8oyWora0VaYjuCmm4Ohh+v79NPS6rCZZCYEbYEg83ERvGFiZJDsWJTsiUoLYCJbHN4HbaTbXmeFKexvOmjxKCugRua4v8XkEkGFLQS0yifp9bTK5DCnqJySWmAgtf2YaFr2wTTf9KhvaFz+NEMBxNaczJ82uNLORUHwyjPhgWShj98vN0lPX6YFgUodKY8mg+j1NHNqDrMRtzY5UtKDfkcdpRXJjMqXmcNridye7SJUP7iryfkQBiNi4uTHajpgUDZwTSuVMdmAqttIJqn7Ym6prMdAgpJ2dG7gCga+6Z7m6EIjHmkariO6PFi/E5IGKH8dkwW3gkffgkojFV3Auq58q0v/YA5bi6M6Th6mC0tcdlVZQcthCiVQC8vuMAQpE43E479hz8UowrfjhJp5zOu+Dy+bUgX6P582Q8GZIRA05KUV+nFTCRGWg8a+JQy8JRQPPUaKLaXhMQ3lhNQ6NuQjFKH9HUxmnjZmwwfo/oMxRy0v7TGmVSfy2irBO77ePqgG5MRAHeVDIdVCQJL+YBsNSxmSJ6PK6KmquC/Dx8uOgi0ZvMn6Cs807VZuMrzh0suhFbsR/pJ+Bmxch8wucSUkY2XigSNyVUGIkfZNA9CSPM/+XHoQUQvVYytC8KfR6dYSS2KfeK+uXnpTAkPU4bvB4nysdphpw8yvJxeiZoJvHlkqF925xRGA6HUV9fL0OFZnjsscdw2mmnwe12Y8KECdi8ebPltlOmTIGiKCn/lZWViW2+/PJLzJkzB6eeeio8Hg9Gjx6NJ554ojmn1uXQ1h6XWZuOWROHWuoPFvm9YqXaGInpxkYFdV6YyndHNGiatCurjuoMiRHcWBhpxGZUfCA1bFRW7BfUbbfTZkq3LvS5U7zHMBOZLRnaV+j+AUnmpMdpZ6oQdqHADujp5kriPJITuF03NiqD0GdoG+PYx1qJ8FBjurGZIjqNzRTuqcDbqNBuNbZiw512cj7Kx/lFMTLR8u+dOQb3zhyT0t6E5xepxYvxnJeu3WXqvdLiIRlWjKEg34XycX4dK1H7bmy63l6c9chrFckr2l4TSGFINkXjYh/5LgciMRU2RavhKh1WIAy61bPGVfLbmlFIc4n0uAx4/vnnMXfuXCxatAhbtmzBuHHjMH36dBw8eNB0+5deegm1tbXiv23btsFut+Pyyy8X28ydOxevvvoqVqxYgZ07d+InP/kJ5syZg4qKiuZfWRfBKaecgkOHDrV7m+2oRYJrf/0JEIvW63GIsd/QdJI/fE476x9lVzBr4lBW+2LXhaVo8uLg1Go+Bhjt3s7lhezIdzlYWCz5M1agTZycbk19lg4eCwsGG73Gg48bdx/SUe3J+xsx4CRduAqACLv1djsxYsBJjFyg/atA+xyNC/KdOmUQGo9NhA09TjvKx/nF9vQ6hdDS9bTiXmrpsALduRrP2wgudgsg43j2lOG6omk+tqpd4sfi+UTyplwOmygx4OcMQCx+MtVpzZ4yXHhkRBIhEoiVwTarPeNeGIWYB3ndIpxMnhnlOnOht/MwdFvVc5Hh6u4eV850+AkTJuDcc8/Fo48+CkCjdA8ePBg//OEPMX/+/Iyff/jhh3HPPfegtrYW+fn5AIAxY8bgyiuvxMKFC8V2JSUluPjii3HfffdldV5dkQ4PAP/85z9xySWXYP/+/Rg8eHC7Hff8B9ehuqExQVXWmGtkGKx+EDPG+XUU32RH5CSl2eO0oSDfhYPHkol9zhScMS614y21zNAmf7sY3ztzjOhYzEEMt3SeBamnZ2ofz2n5fMyp9gqSOThO66bXc4FRrojvI10ZglU3Yy6CrG9Fk6SCE5tPRbI1CO/UzFtzhKNxuBy2tG1DXt9Rh8ZIPCF6q6lVlI/zC8PkS7RCMSsDIKN8JNiEfvl52F4TENR4Y2lBydC+gg1KHYzpnliVLgSbokIsON/lSGkzQhJh9N1alRMYywUGeT26Y1m1r0kHq7ILY5fwluCll17Cd77zHRw6dAgnn3xyq+23PdBmdPimpiZUVlZiwYIF4jWbzYapU6finXfeyWofy5cvx1VXXSWMFgCcd955qKiowE033QS/348NGzZg9+7d+N3vfpfL6XVJkEtfW1vbpobLqOlHYb64mmSuZZqEzVbUCvSadS6HPaV+i7d7pxU6pw+Tx6Qa/r9sw16xb4/TzqSN7EIPzkifdjlsYhIjSjagrcKdNk23jhtYq35RvIbLzY4NdkR3Fv2syDBwzT4gafTdzMDo960ZGJrwG5nx5ZR4rtzPnWij2KzKXufHU6Ev1tXuVcyyVQh9f4U+D44EmxBKeD+UE/QlrpWKyDWVkojueHTO9Duh77dffp5Oo3L11pqUfB3dT/reyGMy6lnSPaLzXbRyG4Bkix+KEMTUVNV68q5oQTfI6xalF+SFmYGeMWO7G3ru0pVQtBZqamrgdDrRr1+/Vt1vZ0NOhuvw4cOIxWIYOHCg7vWBAwfik08+yfj5zZs3Y9u2bVi+fLnu9d///ve49dZbceqpp8LhcMBms+Gpp57C5MmTLfcVDocRDofF38eOHcvlUjoNyKVvS4IGkHxgF76yTST8AX1tEF+xG2FXkg84/QtoITxqCsiLLY19lQhFfi8+rg6gPtiEEQNOQk1DyOARJc+NJ8qNBcBkmHjOQwFwV1mycR7/nMthF2oRxgaPZhp7ZvugY/N9ZDNOrvSbRKt6M2PJNe6mjR6I0mEFWFyxPaW+itd2FfrcqA00Iq4mPTFiLZKgLbUH4Z4Lb3FiVWNGhpqPjYK9dYGQTlCWel/l6omGIjFGconpiuNjqtZzjIw6N7R2ResxR4sHEgV22hWdygYZr7xEPpDneI0aiYDe2NYFGgV7cntNAMs2NKXUbi1auU14a2SQs63tak0QFb47q2YA7cwqXL58OcaOHYvS0lLd67///e+xadMmVFRUoLKyEv/93/+NO+64A2+88YblvpYsWQKv1yv+a88wW2vi5JNPhsPhaHPDRROTCv2KndOtM9UA8TopPjaj/lpNXCRwyicqDt4SwyxRTvkRAuU8aEx5B6vP8aQ5Lz425jwWrdwmPse9keYWjxrzg1bnx+/lsg17TWuqAoxpGQzHxGKD3ucySrw9CPdcmqIxkefj/ESCsVUIjc0o5ICi0/ITElpsgURDn8cJjzNJepkxLpmL4ZqKFJqdMS5VdssYHubeKDFLuZfvYcQYIte4GRuR6zGShw5oIW4zHUkjY9CMdAOk1nBZsQ5bEz2BCg/kaLhOPvlk2O12HDhwQPf6gQMHhGCsFYLBIJ577jncfPPNutdDoRB+8Ytf4KGHHkJ5eTmKi4sxZ84cXHnllVi6dKnl/hYsWIBAICD+++KLL3K5lE4Dm82GYcOGYdcu87qq1kPyQeZrsXTdjTnGFnotSRvGB9JqredJKFeYse48ifouLhSbThiVawAmBXvtKQn7dMWf5IUC0Bknrkhv1OUDrLXnrMbp+okZE/apVGxtcs1zJGuwaKI13mervmlWvc54t2Qjg5IbChr7PE7YFCWF5QeoOi0/Mpp8gRRPeESTR/bHiAEnAYD4l747IrlQ3RnpORJZg5NazAguFGJNQjEl1xxvjCTCyFqol7xFBVodnCiDUCHOYePuQ8hLdDzYvK8ewXAUS9fuwuZ99SlkGIIZS9YMrUnU+OSTTzB8ePuo0HckcgoV5uXloaSkBOvWrcPMmTMBaOSMdevWYc6cOWk/++KLLyIcDmPWrFm61yORCCKRCGw2vQ212+2Ix62T6i6XS7QF6eooKSlBZWVlmx6DckXGpD+fXDQZnrwUmaZCn9YKwqzdeU1DY0ofsPtX7xR5FB5+nDZ6IGZNHKoLhwEQY95igjwis1DKik1VulAS99wop6FAM0zBsNZ99/UddSl5PjJUwXBSOYKKYQkbdx8SDMxoXMWPnv0ANQ2hlJwMjZ12BcGw1i6lpiGE+mBY3A8FmngxnTtNZrwUgDT06DqM5BECD2Eaw5nGliNAcgI265XF5bf4d8JfJ5ID3TPe2t7lsItcXzp2HV0zlRx8XB3Q1Z3ROBpTEY+rCEXiIheoIJkba4zExJi8SNpm2uiB4vooVEi/D5LuMrIl+fnx31IkrqIhFNFJm4Uicd1vj5+flfYnIZ08VGv064pGo/jwww91jO3uipy1CufOnYvrr78e48ePR2lpKR5++GEEg0HceOONAIDrrrsOhYWFWLJkie5zy5cvx8yZM1OShn369MEFF1yAn/70p/B4PBg6dCjefPNN/OUvf8FDDz3UgkvrOigpKUFFRQVisRjsdnvmDzQDk0f2R4XJg8UZarQNAF3fqLpAI4r8XlPRXIfJKj8aSy44jDJDQCrJw2pFmu5Bp91yooGRdMDzFI2RuCV1mZcGRONqih4gD4tywoDZpBuJaZPdcabZx8+JPg9AZyD5mF8Hz9MYc0k8J2c1TpfH43VM9D1Yjen7IYaoTVFgs2nXG47GU3KXAZMwMKBvPGpTNImomKpikDepi2imycj/Mja6rA9q37PbadddH4XHSaORE0VsNm2BYUQ28Qd+/HSfo6L7huoAhhT00jVnBdDqRI0dO3agsbERJSUlLdpPV0DOhuvKK6/EoUOHcM8996Curg5nn302Xn31VUHY2L9/f4r3tGvXLrz11lt47bXXTPf53HPPYcGCBfje976H+vp6DB06FPfffz9uu+22ZlxS10NJSQlOnDiBTz75BEVFRW1yDN7UkT9k+S6HoDDr2ILs4Szye00LhwHAYdd/1ys2VaVMOgQqOhWeToK6DJiLw1qxsD4/HBRGlIfA9AK9evh9blPxW3qPWJZ8DGiTz87aY4IgUZtI0vd2O3CsMdlwUttfUkSYdw7m56jvKMyNfnKsV0Y33ybT5JdO9LcuoLHj9tefEGFJztAzGwP6hQD/jkORGBPXTY7Nyit4aDquArHEj4zfc9Jq1Bkr9jc3GvpGl6ou1Lrqo2RDTiOsfifZINsCIv7MbK8JZL1QaC4qKyuhKArOOeeczBt3cci2Jp0AgUAAPp8PTz/9NK677ro2OcZZC19FKKL1aIrG41qti13BxWNOEYzAf26rNX2gtToW1uYdyUnE47ThrrLRIgTH80ZG3DdzjOX7nJ5OArcUevE47UIlnIfOjOfC66DMroFYX8aJ2KoeLNN7fBvA+nP8HPnrxvo3GhOMjEnensWq/sg4XnxZ8p6bGSD+mnFMtViUWyTKudEgeZx2TBs9UFffRaFE/n3ZlSSr1J1ogcLp/BweQ5lAcaFXLCAKGcOSRwzoM8baMavvLVs4bQriqoqYqi2OBvR26Qyt8XyJ9WnG7mzLVidz5szBG2+8kRXDuzMilzlcahV2Ani9XowYMaJN81wUvovG4jo1Vx62s1qFxlT9ypZv5XLYsXTtLiENZBZ+ISzbsFcXEuNkAqr/Wr21RvzLw4E8vEasteJCr0jMFxd6dV2XnTZFx/Aa5HXrclp88qYmhMZmj6SkQIw53pST6zQaZaV8HqcgU3BxW06y4IZKO09FeClUY2RUOOcis5wibhzTPSbZJMophaMxnaSVEfz9cDRmKXZbPs4PL5O9aorGUDqsQOgD7jl43FS9Y/FlYwSrtCDfhSPBJtPzMFPS314TMGVY6kWIVXHd/Ps1+1kbA9zpyOORuCr2wYWFFSQXLS5HUisxwLprU/7tSLDJUniXM11bgsrKyh4RJgSk4eo0aGuCBicY8DGF6NL1cbIr+gmC92SaPLK/bh9WYUJionH1BwoTkSI9eXa0Mjc7HqCYds/lCXpAC4FyeSmabAB9mMvjtLPJNE+M43EV86aPEvRxl8MujqEVqCYN8JFgE5uENcUGsxYdxjEZrcWXjRETtdfjRH6eFsF3OWym7VmITq4iyaKjMRlBQjjKvTiFeafaMYYU9BK5GFJ1p3E6sduGUEQcm0gZ9DsIReJZdUQOhqNJqTBbUjYsz2EX3bMJ1C6FjxXorzW9+dFT1f1s4cHbs9iVJHWfCuN5Z2y+SFEULVdG18Vzd3RfwtF4u7Q6iUaj+Oijj6ThkmhflJSU4IMPPkAsZq7a3hKs2FQlSBR+n1uMxxZ6dZ4YPXA2Re8NxVSkFC0TNu4+BN5YzwxOmyL05giRuCqIG3E1SUGvbtDqhJJagqk9urih5ONlG/YKYzBv+ihdLyxuCPV8En1PJd4ChYc1rQy7Aljq3lmphNPY53Hq2H52JbVvGh/T59OJ7Wq1SjadJ2A2bmBeAX0vRpJKJrFb7ntT41C6v2YTtJnxiyeOGYmrYhyKxFJCcfx9Gqsm15fOdHHPi0cQuBcVU5OEIlpkRWKqWKhwLyquJnNlRuo/oP9O2rrVyc6dOxEKhaThkmhfcIJGa8OqkPVIsEnnfVG2U1UhVshUO9TIal1obFf0xsrobNGEku9yiBogMzhtiqg1os/xolJjjy7uhfAxTeyLLxuDWROH6v7mYSl9VlcReZaPqwOWORdA1bVpobCd22nXteIwtuVIB6t2KunU8pE4Phk947jQ50kR1zUb8/5aVFzLe6/5fW5NjcOuiLYjKqjHVjyFMUmtcAAiXWjnTLT+jbsPWYramoVpaeFEvwn+Po3JoNP5Tx7ZH95Eno7Cu8lvWe+588VLkd8rFDWcNkWEco11ccb8JEzOVYF2H43fSTatTlpSz9WTiBmANFydBl/5ylcAoE3ChVypnT/8s6cMF60oysf5dWPqTEsSSpR/cTttYrz4sjHg4Rn+mBcXekUh8bzpo7R2EBZhxIvHnoJZE4fq1AmMuSsyQASqUeIeCWHzvnqc/+B6XaFoydC+YoLx+9xizKn7HIU+tzDKTruCxkQPLhVJJXRt8k7mhWIqhMGKqVo9292JppV8TG1CrIqcjQXPBKLzU3uNBtZqg8a8dYfRkPIi6yPBJiy+bAxKhxWgsuqoMO7Uo6su0Cjo/ZmgLTRiKa/RfUh6efrzCLBeZtTtmsK0kYTHRx6pUfVjf/0JU++FfgfaWSd/kSqA/DyHOBf6KSrQwszCu4qraIrG4HLYGenDpjP686aPEj3nBvR2Y0Aft+44R4JNKd9JNq1OclGaN6KyshIjR47scsS05kIark4Cr9eLM844o00MF+VGeA6nIBESe31HHeIq8PnhoBgbQdqEKvTSUbAcaQ8vf1DTxfLNWt6Th2TcDw83mUkkGQkeDSHNq1m9tUZ4lzUNSa+Th8f4+dcFGpMeRFzVvUdkCLPP8XneqsbMuN3G3Yd0HiYfA8liY6twpJn6hhlpI5Aobg4k9PSMPbf4/sk7LjTkgviY+mjxhQZ5IkV+r8idFfm9wqBQ3ofuG90GXq9G3pSRzWm8FrN7oQ+9JW+y06Zg3vRROu+SVEHoGGbfjfbdxVM8aTPvyawxZaYcVjoVlVy8r55EzACk4epUaCuChlWehTd75CGg1VtrdKs/Y+0NgYRurY7JkS6Wb+yVlE7Lzfg6/5se/EFeNxToE/o8T8dzdNz74uEfqs3iYyAZYsrkh/g8Dl3ok4+NvbGsCBSAYkpyyKRvyIkTRgJHUzRmaiyIaRkIRfD6jjph2OuDTcJv4blGh00R+6qsOiryPvR74GHX7TUBXR2hmfyYy5Fs1rm//kRKeFNhn6KxWWiWh5y57iBdD4VezZiPFCKk13iokb47IqKYeU9mjSkz5bCs8l65eF+kmNGTDFfOBcgSbYfx48dj1apVra6gYZRkovEL730hinB5v6dBXrdYxVI7BzOko1YDWv8hMjDpaPKb99Vj6dpdok3J/at3JN5RRGsRAKJlBNeJowmRZKRiKpJhrkS4ieCwazU/9C+QFI0F9Ml4LtaqS+RnjpwBAI43RnXK5AQzlRArz4zGB481CgZmv/w8HAlSVwRNgir5/aiC1MBDoHyfNkWrSVKheXZkBOh+GHNXRuV+yjXyVja8BYzYUrFWt+AglYxjjRGhPUgKIVQU35BoQxJJsB5pzD1fCs02Jcol9hz8MuGVJZX0qV8XXZfxGHQvmqLxBOkoee35LrsIsx481oizF7+ma1VTOqxA57HyIm7+HBjbnqRrdZKtmkZPI2YAsgC5U+HNN9/ElClT8P7777fqj9Co0UdjnmfhoJUlb6aXKwp9ycZ7ACyNHz9eun3RPqwKgmlMOTKu28d1//JdDl0TwSJ/shGjPvmeLCjlY1tiQnbYFDjsNkRjyRKAQqa6wRs8ckUPp13R0c5zBfc8mguy5aphnOmYCvRNJ4Fkmxdt4REX98noIdExzAqtzY4F6OXIcoHVPTJ73VgETpqL6YrZjdC8/NTfu9WYP1+Zts2m0eSTTz6JO+64A/X19V127gNkAXKXxXnnnQev14t//OMfrbpfXiDMx7z9BOUynDZFF6vnYTIeOuG5DyOcdsW09YPptjZFV29lhJFWbiwIJqJJWbFfMLlKhxUIQkjpsAId64yMNy+IBZJKFDSeNnogfAlF9GmjB+ro+DZF81AK8vME+1KBFlajMYXOgNRSAjIC9KoCaveRVH03hqkIvAWHlRo8h9Nw481CbhyUt+LgKvHGNi8ARAsYl8MmWImcoUiEHgIfc1ae8fp4/pGfl9k5cvD7TQXhdK2p9zUZmp02epDutzJjnF/8vooTHQiSLEebrrYrXdmDcWzV9aC5bU9WrVqFr33ta13aaOUKabg6EZxOJy6++GJUVFS06n6tGGs0acfVZJ1KXFWx5+CXIqRVY6inAah2qQBW86aZ8K4VFT6uqvj8cFD87WGhJKddEYaBkGSWxbHn4JeCkVY6rEBHL+c5ArMcEOWHKC9WVuwXDMWyYn9KTomHyMiz45MKhd5orK/7Sl479YPyJhQoaDxv+ihMGz1IsDppXJwwANq9sWHa6EHCIJx1Sh8xLmaGgiZZI7gR48QIvfHS/6VAK1LOS9RIGRmQyzbsTREBtilJpp5NAUYM6C32FzV4UPl5DsQTQR9twTAoxfjReY0V18XHqeBeWk1DIyaP7A93Yp8jBvTWqewDqtiPsfygdFgB/Alm7RXnDobf52Esx7gIf/JFSktgljfLpKYRDAbxxhtvYMaMGa1yDl0FMlTYyfDss8/immuuwRdffIFTTz21VfZ59uLXTLXo5k0fJcKFXFOOVqc+j1OnSM5X6OnCe0adOyB9qJCD59oIfB9W50nXpUCb+LPpWEzK6WbhGWodb9SV4+PSYQW63BzdHZcjmachj4PG5eP8pi1LCn0eEcI0Xg+/zTyEZeU9ZQr/WW3Lz5WDf9c+jxOBUERsXz7Or2vnwcOKtF/+eeOxvYaQHL8nfGw8D+PvL10I0ur4xs8Znw+z37DV789s21xDhZlCgmZ45ZVX8K1vfQuffvppl+/DJUOFXRjf/OY34XA4sGrVqlbbp5USA6+d4oWcFK7Jd9lFga+CpEyO067olCh4+MWmkLHQpodgOIKDx5JeG0kNAVoYp9Dn0YUdjR1oNXmlKEqG9hWGNM+hsb+I5cZZel6P07Jbshms6MiZCoJf33GAhcjsaEoYKpfDjnyXQ3gWKqAbv77jgI7Jl6mWy8x7JVipmVA40q6Y09nHFnqFAOzYQq8o2jU2lKRQKYVhaWLnnpqRaGJsPkmhMfp8+Ti/8KrdiYaQCpLfNaCv/eLGiYetKZwNJD1hq7wW/5zDoGHJQ4WTR/YXFP4hBb1Mfxua92lPGF1H2m2txiRr1hrSTxUVFRg9enSXN1q5QhquToa+ffvia1/7WqsaLivK7YpNVTplAxrTiru6oVEwslQk9f7icVWnRMFX6Aq0kAeRFBpCUV3o5ngjZxcq+Pf8C0WYkowQzSuk2dcQiogJUmODxdCU0EVUE9fHJxwu4QRYdyy2uv5sZHgaI7GUCZbG1Q0hXQkBH3OGH78v+louRYwd9qSMEU349De1/zAbUw6PtrYryfvJtRWPBJtEgfe86aPEeNrogch3OTBv+iiUDisQ4/RU/lQlC6qton1XVh0ViyHeAoUXOpOx5NcK6PUmOWWetuXG1a5of6vQyzs57DZBrwe075HGlVVHhfzV9pqA6XOzvSYgRJ/p2dhaHRALGipOTyfzREzWlko/xWIx/OMf/0B5eXnOn+3qkIarE2LGjBlYt24dvvzyy1bZn9UKUJ+fiIsxJct9Hocuv0JelvZvcqLhyfUiv1dXNMknHmLxUcKb+nMRpdnttAlVeJJ3MvNC6Dh8P3zC4WMqyDUrVrW6/pKhfXWG0Gzs9TigQBOEJa+Bxkn5JDvL1dh1yX3jfQOSYTWXw2Y5bghFUrwaszF5m1z2ioxSNsn/jbsPmRp7o4fFJ34gfT0SjcPRuKmSCwBTId9styXjarxe7o3SPeReI4U2++XnidcHed1CcaY+2IR++Xni90vfJY8eUE4zG2Hd1vK43n33XRw6dKjH5bcAabg6JcrLy9HU1GTZeDNXmK36Nu4+pAvhmXkCxxujaEo8kE3RuG6VzkVQuedwJNikK5rkq/G4Cp1yx6yJQ7F07S6xTTSWbB9RVuxPhPlo0lHE2KjoDiTJH4O8bjEu8ntNhV05MSOJ5JkaV95m44ZQVEg+GZso0sqcjA4fc607CiHS4sDoPfDx4svG6MK8ZiFePib1BbPcHh+TGomZcUk3uXoSSvRkMMnbMerwWYnz0nfHv0czw0r3LZttaTFmvEYi+FBe14zhSt8T/ZTrAo2ClMO/1yPBJhQkDBy1L1HEHrRxJmHd1vK4Kioq0L9/f0yYMCHnz3Z1SHJGJ8WYMWMwfvx4/PnPf27xvn707Acp5II8R2oSnlCcaASYjpiwcfdBNISiKQl9LVFtT1H3BrT816XFGjmhyO/FkWATDh5vNNXD8zjtuKvsLNz9yjYAqTU9MxgBA4ApYYEUEOh4xP6aPLI/KquO6uq5ePPFsmK/rininoPH0RiJo3xc8h4Y92dVN2ZFNOAkFD6m6+T3VYGeBGHVgNL4XdC9oHmaPLc8h1aXRt6g1tjRjhEDThJjQBUFunR/M91Po8Hg6vr8e9HyQ86U+2Y2piLfdPebxrxwmc6ZX6NWe6ctgKj+zpmwvtSUEtDGxODkv4MjwSbdNdYHm8R9NH5XZveDxrwAmb+faz+uoqIiTJgwAX/84x9z+lxnRS5zuFTO6KSYMWMGnnrqqVZR0TCu6mKqlsyOxjQxU4qkxNXU8BWJsFKXVtoPrTYBvbYeaQOaQVWT50IrWCuEIjGd52YsRN24+5AwBJwc4jAYOMo3kJirXYGoQ/IlyAPGCYSksAp9HiE8W+jz4JGrzwEA8W86rN5aI/JBfFyRaCdv1UfKTMFCTeyDrspK2UI1+Zv/CyCRw4yJ16lIuima9BybojH0djtNFzbp7id9X+S98dAu/15Uti/6HsgbpvtFv43jzIMlJp7VtgR+XPKk1MS9Nf6O+KKJFgxK4j5yxfs9B79EUzSGzfvqAWiqFlpX8BjztZKEFWMZR2vj008/xY4dO3D//fe36XE6K2SosJNixowZOHz4MDZt2tTifZkVNk4e2V/Uz8TVpJRRJK5ia3VAJJp5GInvx82YYTS2KjImjC306uL86eC0K2nj/jwnxSd+hz35k+ZNLo3N/ZTEPsxWvc0tBCVYkWF4fzCznlr6K0nNHZq9blxo5FqUTHkaHmId5HWLezSkoJcYp7uf5mHFZGiXfy+AvpdasCkqireddkWcB29fQnkhLeRovm3y2pNhZbp3vD4wHSjnZVf0ivdExlm9tUYYS4oqGEOFAEzzg0YR6JYowq9atQoulwvTpk3L6XPdBdLj6qQoLS3FgAEDUFFRgfPPP79F+zLTKjx78WumHo8WxtH03Yr8Xlxx7mAxqW/eV4+6QAgvvPeF0MIbMeAkANrK3aYkuxqbgdhsFOdPh/w8hy7ub6zvcjlsaEqIxFqNtc9r56O/VlXkLoxeQrpwTTY5IwCWmnU0pvAWH1MojU6TvEgKW/EOz6S1yMcEYhcaQ4R87LAlJajIc65paBTH5mPuGRMjMBSJ6ZolAqpOg5HGtC1nEfLvgBCJqYgkGqjG48k8J+W1eF6In49xWwr9Go9HeSmqJbQK4Yr7GVOR57DD5VDQEIomQsnaM0ELruqGRkNNmHaHqZjc7LdBz88grxt1gcacNQk5Kioq8I1vfAP5+fk5fa67QOa4OjFuvvlmvP3229i5c2er7/ushf+0zHERaPKk/A4vNG0OZrAckRZmsTZe980cI4RzeX6NcmNkCADNK9ByMzaMGNBbjO8qGy3yLFwvkGsPFrMcFuUxeO6L8kHGXE+mscdpFwbUaqLkY9oegPgb0CbfbMZOuyIm8nTHM2o10vWZ6Sma3U8gNU9D99h4/YCakiekkKFxWzI2VvkrypsW+jTB3sZILLHfLy3HdF35eQ5MHtk/EfrT7hvP540YcJKpXqUZePmDAs3z5Pc2Xa6Kit1bWoBcX1+PAQMG4LHHHsMPfvCDrD7TFSBzXN0EM2fOxB//+Ed89NFHGDduXCvvXXsC06kRcMO2emuNmHT0K80kzFQvOHiOLZ3HNWOcX7DdjLm5I8Em/Hv+hTj/wfVoCEVQ6POw3ExcN6YJZNmGvbrJmjwBQL+CpzEFmng+KFc0Jmp9zOq8zMZ0TuQZca8hm3F+nsNypW8cL127S+ft0X7o/uTnaU0QC/JdjMnnEvfTbGI2Hovut/Ee57scYoKnbY1kBRoTI5TEngHNG7Ql2Iz760+I1ipm3yOgeXMNoQhWidyiORMUSBVoTi6I7CjId6K6oVGEbLdWBzC2UB+RyKbYvTU8rr///e+Ix+O49NJLs/5Md4PMcXVifPOb34Tf78eyZctafd80WfFctVGIm2dLyor9KQlnu6IZmUKfB/fNHCNW5GagvAdv2W52HLuSJD/wQmJjUTHPpfD6Mj5esalKNEjktVbUa0kB5T+SY2NNFJ2TsdtyprGxfXs240KfRzQ4zOYYxjGHle6dWXkAwUjB59+XVYt5fo+pPQ2/38b6q3nTR2XdBdjsO3bYFFPDz9U5yor9KTqGPAxr/N4pf0YyXoQrzh2Me2eOQQHLq31cHcCeg8cB5K5R2Bp0eFVV8fjjj6OsrAyFhYU5Hb87QYYKOzkWL16M3/72t6iurobX6838gSxB+Roecps8sj9jvenBQ3dWVOjN++pNP0/hGq7fx1lmxQbKOWkAmrET7Qqwd0mZpcag2RhIpclbhdLC0ZguvBWKxOExhB3bIlRIoSaiWOcamqTPZwo7pSuNyLbNhtU9NvPeM4XDyPAB+tAgeY/0OfLgjGHF13ccSAkV+ll7GToPZ6INDXnCxtwWvUb3ENCfO//9mG2fTaiPP3PNpcNv2rQJX/3qV7FmzRpcfPHFGbfvSpChwm6E73//+7j33nvxv//7v5gzZ06bH4/qiJw2BQP6uMQEQCQGCtfRA0sMq2Ub9opwjhEUeiKoSC1apomzdFiBZZ8wQFsdn//geuENBMNR0dbeOM53aTkgXqtDsArdUaiTh5tCkTiWrt0lFD7IuKUb875U6Y7HxzRBZgt+vGA4ikBiYu+Xn4fhC1abigLT97V6aw16u52IqVoO04pM4nHa097jIQW9xETP85ZW4TAjwYWLPNO50f3gTRbJoLkcNvHbq6w6KkKFlKPS7mMyDC1ChnEV+S4tHGvMU9GvIhiO6qj0/Nz5tfl92rWRl5ttqM9IkuKvZ4vHH38cw4YNw/Tp07P+THeE9Li6AC6//HJs374d27dvh2LRHiRXGBPFgLniNk3kxRaJcHqIy4r9+PxwUEwgRg+Hr2SNuG/mGF0NVTZNJ41FwzQ2U6Pn71NuZ9VHNSlFt9NGD0whBoSj8ZQckNlxjWNO6sg295RuNW425nm7uIqMOTVqdkjwGZTeeRGtUSXeiqFo9fsx87isvE1+7/i58W0JRrX/F977AlurAyj0uXHweBiRmKrremz8TSuA7vfrZkQdI4zfaa7eZGvj8OHDKCwsxH333Yef/vSnbXqsjoD0uLoZ7rjjDnz9618XHZJbA7SaJZqv067grEF9dCtXPnFZJb+JOq1N8skJgLdup8mGVDCM4CvRzfvq0xou6lhM9UNAai0RAN21GWn9yzbsNZVrIjHYZRv2onRYgSi6zsaIZJJVMpYjZBpnsy33XmgC5+QBs7FmQBwIhKIIR5MsP+0eJAuejSQb1WJc5E8aAT4287h47R0fk8ZjoDoAW8IIG7flnhJHUuxZwYDe7sRvJ6k0z9vy0Lnri9+TV2NUZ6FtCn0e9MvP07EQt9cEWkRnbw7++Mc/QlEU3HTTTW1+rM4Oabi6AC644AKcddZZePzxx1vNcFFSmFamRsV3QP8gkxEwjokOrE0uyQmCEz2obswMxskoU5KanyPVktHYyECkc9xz8LjIowCacatpCKV4XNzjM9Z35VKAbES29V/c0GX7GQLlHDlhwDg+3qh5NHRfjJ4G/76Nk7gVPk4sdFTDmBMQADIC5uaPLyDMDAqvxVu0cpvwghat3CbCoWaeaKHPY2jombpnY/F6JK7dk0IWDqTwtQotVEmGj19bc7QGc0EsFsMTTzyBq666Cv369WvTY3UFyFBhF8Gjjz6KO++8E1VVVfD7rVvdZwuaGAEV1Q2NKC704rST8y3JGZngtGnFxzTX0cRHxIaFr2yzrAHjoUIrggegGck+hpASQQFEXykKBZodLx25gZNTjLp1Loc9p/Cg1THMxkYih7HWKpvjkVdirPMiNXSP0y70+XgozWlSV5Xu3OjeK0guWsxQXJidNiR9bxSaDTZFEzVkyRCexxDOM/6tQMs71TQ0alqEqlaITRqF6SY4Y0iSh04pFGp1/7W2Mc3XGswFa9asQVlZGd59912Ulpa22XE6ErnM4dJwdREEAgEUFhbiZz/7Ge65556cP8/ZW7w+hSYzj9OGgnxX1p2KM2FGotMv1cWYGRsAuuNy9pgZaDIx8wyKE8XDtJ/6YFh4gHz7dIaEi+CaHRtIzfGY1a5R4TYxEq3GZAzNJkkj683s2Px4ZqK7dAwC3Veer+GFs0Bmb5BCkrTQMeYK6bzM8j9WXiSf8Cn3amVQgOy9QWN+Syeoa1fgsClwJcojKLcZDGv948wKkQt9bhw8Fha/Nx5+puvhOcrWNGSXXnop6urq8N5777VanruzQXZA7obwer2YNWsWnnzySUQi5hN7OlD9TkMogq3VybwOTQChSFz0H6LwHXXBzQZcIs9pU7KuT7mrbHTWeoDUcZZPWjQmNQ0znTyqP1NAensxpsMXE69T/Y/TrsDjtCX+swvDYtbdlwyDXUmqXbgcdvE6NxxGWNV8kZYkNTykqzWO6bjU/ddsTDVb86aPMu1obKz/AqCrx+JjQB+SpLYfvOWIjwkX83Yj5z+4HgBEOBaAqOfioM/Mmz4Kiy8bI8Ze9juMmhgtPpU7bdr3l+ew6ZpLDujtxoA+bjEuyHehIRRBZdVR5Ls05ivt28wuVjc0it/b9ppARh3C1sK+ffuwZs0a3H777d3WaOUKabi6EGbPno2amhpUVFTk/Fl9L6WkwCs9BjYFeH1HnS7PEIrEkO/SjBeFzshAOe2KbrLgfjut4GkCM5scOUjDbfO+ekFhN0NDKGr6ul3RJjxSLtdWz3RCquiq7HbasHH3IYQSHsnH1QExphW3TQHOGtRHGJ9poweKflNkTPiEz4uIRww4yaIwmO6UohvT+e45+CWON0Z0nZ65l2plvDxOe1ZFzbQgMOtobFb8u3prjVjk8LFxMrYSbzbCqilltti8r14YTruiEX+M4HYmGtekpkKROI43RnT3gZ+zscmjAm3fFHqkRYzTpv3WiwuTRctlxX7T6y8r9ucsypwJTzzxBLxeL6666qpW22dXhwwVdjFMmjQJLpcL69aty+lzFILh4DRqPiEaQeEwp11BLK7lsQp9bgTDMQTDUZFPMBYRUxhlccV2y9AOyfkQLdoqVKfldJJhOWKfaftwY/aUM3SsRa69x3MzPMzE92EVKuTnZNZDyyx0ZwxBErPNKmxodq+twoNm31O6UKICrT8UbyHCz436a/GaL840HdBbq+Uj1qmRWWcsRjfrRUbfRXGhFwAEfT0YjonPGXur8VBjJv3ATDDmoyicZ5ZzM4MxDwek0vKN+26tUGFjYyNOPfVUXHfddXjooYdavL/ODJnj6saG669//Su+973vYdu2bSgqKsr6c8YcFzWKpNqr4iwS2enADYIxv2GVN6OaGjqu0TBoxIJoWqNKMB6Hf4bIC6SaYLbNjHF+/PPj2hTyQqHPXAw4HTmhJeQMjwUpgY+zuR/GczXLDzltCuJqqjAvIRudRbPjkHdrNMBWn+OGycxIaAoZdSmKJkDqeTttmssUiamWuU1+nknv1QZAsazpMp6nlYIIr/Fqjbqup59+GjfccAN27dqFkSNHtnh/nRkyx9WN8Z3vfAdDhgzBL3/5y5w/S+GhK84djEFeTaWC5yxymQyNbZ+4R2XV7p2DJg2iQQOpDRKPN2rGgya9dNF9np8jg0j5qckj+4t26xTyLC70YmzCAygu9KKy6qi4BjouoBcD5r2uxhZ6RU7M47SjuFCvx0et5bMZ31V2lsjnmPXpMo7HsmMVpxkTHOy8jflBs7oq6tFl1H6klh6DvG7TfmoUwtP6syliTPeZ74+OQfsjqACON0aEh0gU/oJ8F+6dOQZXnDsYBfkuoZFJuTAyipG4injiGkkQ2gi6H8bfe0FCB9MMCpJ9vRSk/sZ5nzl6zajpSONs0dTUhF/96lcoLy/v9kYrVzTLcD322GM47bTT4Ha7MWHCBGzevNly2ylTpkBRlJT/ysrKxDZm7yuKgt/+9rfNOb1uDZfLhcWLF+Nvf/sb3nvvvaw/Z5ZINuYZaH7LJv2bTpA3G/HQJIsuOYFxY1joc6Os2J8gPdgyhoqI/h5TkzVFLocNKpJq44U+jyBqUG8wGpPhM4rC0jkAmuGn8ZFgkyk5gSa/bFXajQQIyvHZbIoYF/m9GFLQS4z5sdKNCfkuh/h+KEdDY7oelyM5Jm9zf/0JoZx+JNgkjHhdoNFU3T8Zu1FZE0ZV7IPvz2pxACTrAmls1nhx9dYa3T3l3xkf+32aURxS0EvcQ45kHaGC2VOGm/72PU6bWGTx35RZw1Bev5bNM5cO/+///T/s27evx3Y5ToecDdfzzz+PuXPnYtGiRdiyZQvGjRuH6dOn4+DBg6bbv/TSS6itrRX/bdu2DXa7HZdffrnYhr9fW1srKsS/853vNP/KujGuvfZajB49GvPnz0e2kV6rrr7cA6BVvTvBqEsHvmIG9KtX6lKrIKnmboQzofLNCRd6Y6jgkavPwd4lZbirbDQKfR6Uj/MLr6fQ5xZjp01JWT3T33SdpERO12u8Dxt3HxIFpuR9uhw2lA4r0BEyaGz8PB9bTVjGsRkBgsKSkZgqxttrAmLC314TsDyuccyZg0QwmTZ6kG5svDaauI2hwdlThgsjXlbsF+PCBInBk/Cs6HP8+8hGzd3ncRqMRnIPRX6v7rooHMfvKYWbjd8fGUV+D6PME+NEpM376oV3qP99K2Jbq27PZh6X2fvZKmwEg0H86le/wrXXXouxY8dm9ZmehJxzXBMmTMC5556LRx99FAAQj8cxePBg/PCHP8T8+fMzfv7hhx/GPffcg9raWsvunTNnzsTx48dzIiD0lBwXYeXKlZg5cyZee+21Vmvfffbi11LyL1ZI13srnf4hgeqHzOqtFAD3zhwjkttWCuJU2EpF1HwfgL4IlifTjYr4lVVHBUGE2tpHYqpQtae8CM/bUG7OkagH0qAFQKMxVQgVAxBjKtDm9UBEr48nJnSzMd+e5JqMqh9mY0fi/Om7zEZnkX+/vInktNGDTBs8UuGzMWdH+USrJpE8zwokm4Qm853JfXCCieaJqSn1V0SAyUQ+WfVRja4zgfF3C6T+Xvlv3apgm7Q4rXQM+Tib3Nf999+PX/3qV9i1axdOO+20jNt3B7QZOaOpqQm9evXC3/72N8ycOVO8fv3116OhoQErV67MuI+xY8fiq1/9Kv7whz+Yvn/gwAGceuqpePrpp3HNNddY7iccDiMcDou/jx07hsGDB/cYw6WqKiZNmoTGxka89957sNlanq40KxCtDYRMa1qs4DNJrBPpgbMTLx5zStrC5M8fTIaSOSMyG6NIMBI0AH1y3mwfnExARAEzWnpbgx+rpcc1FvNmW9icSUCXn5vecOmLoJsS+TMr5mY212f8noyGl9TbMzEXeSGzEcbPUtjPeM+M1+tjQsrZiCRnYhseOXIEp59+Om688UY8/PDDGe5M90GbkTMOHz6MWCyGgQMH6l4fOHAg6urqMn5+8+bN2LZtG2655RbLbZ5++mn07t0b3/72t9Pua8mSJfB6veK/wYMHZ3cR3QSKouDBBx/Eli1b8OKLL7bKPs2aD/Zxaw+oFg6yibGRnEE43hjB54eDCIQiCEdj+PxwUFfkDADRmJrWaNkUYMbv38LwBavxo2c/0IUbeSKfwjJWYU1+jlyRg9cB0T4oNDSWkTbGFnpFDRiFwmhsDJUCet1FqzGgD0NZjTmhgo95KI+2txrT9lRcTXVl5H2oICJFcqwgSXDhZATjfSf4fe6U2jWAh24VUwII/86yqak1kkGMeTCzsGp1Q2NKODadgdxff0J46PXBiGmImOfMrJp4moE/W5mwZMkSxONx3HXXXRm37anIyeOqqalBYWEh3n77bXz1q18Vr//sZz/Dm2++iXfffTft53/wgx/gnXfewdatWy23OfPMMzFt2jT8/ve/T7uvnu5xEcrKyrB7927s2LEDTmd2Khe5wEqmx0rpPVtkK9tj5gnw95rjiaTzuPjqOl24h0KLHNnWG1l5U3xsVjNmPBaQHVU9HXVbr32YDIUmyTPWbT342Nh0EbD25LK5N8bPZfK4ivzelNAj76FlVdbAYeaVW33vmpfnSUuH55/Llhq/f/9+jBw5EgsWLMCiRYuy+kx3QZt5XCeffDLsdjsOHDige/3AgQMYNGhQ2s8Gg0E899xzuPnmmy23+de//oVdu3al9cgILpcLffr00f3XE7FkyRLs3bsXy5cvb5P9W7WAt4LRA7LyhtIZLQXJicthVyxX5A67kpYmz+nr3PvyM0IB0bMJ3CtIl2DntG6+ryycB00I1mKcZPnZTceUgzNS1c3GPo/DkiigINlOhMbRxIXTveJkBO59GT0xvl/A3JMzgt8nocZiU4RX6zZ4uHRv6XOcaTnI68aRYBMWXzYGpcMKUFl1FGXFfh1b0Wi0POycSKZq3vRRwosiT9Lqey/ye3MiZ2SLxYsXo0+fPpg7d27Wn+mJaBY5o7S0VHhE8XgcQ4YMwZw5c9KSM/785z/jtttuQ3V1taUs/w033IBt27bh/fffz+WUAPQ8cgbHtddeizfeeAOffvqpJeGltTHiF2tSjE8mL4oX2mbrLXF1i9T9WRNEFCBtEt6qgaGxaNUsN5Qp52ElesvPzaxpo4IkCQFIJaLwHA0XJLYqbLbynPg1Z/IS6R5nInPwe5Ju2+Z4yZmKownZqm3wHCs1CiXyDnVkNvOcrK6rtTyuHTt2YOzYsXj44Yfxwx/+MJdb1C3QpgXIc+fOxVNPPYWnn34aO3fuxOzZsxEMBnHjjTcCAK677josWLAg5XPLly/HzJkzLY3WsWPH8OKLL2blbUno8atf/QpHjhzBI4880m7HpHoojzPpDTjsNt2qGIbxtNEDRa7FavLitUWAtdEC9J6I8bhjC72WRpTOwWxi4wKu5H0pacbkMfG+Ty6HXYjC8jHdK5Wdu3HMBV/NxmSIstHCo3Mz5oPC0bhpzRN5x/w7pXOzChPyf+l1oydHeUKAKXiwL9moe8mhAAiEkosE8s4Ged2wJVzxQp87pSSgyO+1zH/G4yoqq46iIRRBUzQmdBiXbdgrrsHMkzTex2BTtFU9rrvuugtDhgzBrbfemtX2PRk5G64rr7wSS5cuxT333IOzzz4bH374IV599VVB2Ni/fz9qa2t1n9m1axfeeuuttGHC5557Dqqq4uqrr871lHo8hg0bhttuuw2//vWvUV9f3y7HNFN9ALROutwe8DEv1LSxSY8TC6aNHiQmJEAf7lOgTVKANtlNHtlfbOv3uQWJwWlTdE0UOewKUDqsAHkOzcgOKeilqwej5LvTpohGlSpgOY4lDF00FhehK17syscF+U7dcc3GWg1cWDcGgPpgGP3y8xLGwCb2ne54MWaEedNN+p6MRctUL8e/03nTR6UQO+yK/pyL/F5d92tjZ2muQ+ikcChjJToSdXhmxoz/nlQkFzJcqf3g8TDqg01Y+Mo2vPDeFwiGo/g40SWbL27IIJcV+4VRoQL4fvl5OgNT3dAoBJtXb9V6tNmVpIKJdk/VtAXI2XZIAIBNmzbhlVdewb333guXy5Vx+54OqVXYTXDgwAEMHz4ct956a4eJcZ618FVLrTcKgwFa4pzEfY21TgP6uGGlbQjow3yUHDdDujCiLw3FPRuChVm4i7yBbHX9sqGamxESuFYeJwpkOh4/Z6o7shKHpdeIuk3lCFahOON5egy1ZNGYaulhG0OlZkSV5sKuJMV1qQ7QGBLk2xp/T8YwJ13X1kSnZzpH6k1GhJC6QCPKivVC0+nywvF4HFOmTEFDQwM++OAD2O3WHRK6M6RWYQ/EwIEDcffdd+N//ud/MrI72w7WUw2FwWgVSltG4qpYOUfiakpYhXtcPo9DeFiDvG4cPGbNEotaWB+7og/rGanmRDxIB/LKgGQoyu9zZ0WWMAvRZRqb0dONRAGzMZ2b0848yYSnakS61iNmihz0PTjtis5DBvSeXDxubrR46JDCs36mhuL3udN+D9w7K/S5xd+8FMBmU7B07S5s3lcv2sbwkKBd0YwOeWF0nfw1fq/IkyTQdW2vCZh6XNlS4P/whz/gX//6Fx566KEea7RyhfS4uhGi0Si++tWv4ssvv8QHH3wAtztVCLUtQcobZiDSAHVfph8dJy9ok45iuY90q3wjrFbsfIVv1lrEjHiQ7jzMPtdcskQuBIhsxvzcgFRCilGVvika09HJrcY2JUlg4ffZbB+ckm4FK0+Le8ZGD1r/uzEvLjf7vmaM0ytoOOxaFwJSUOGSXUaKP6nWV3xUo9s/74aca1uTzz//HGPHjsU111yDJ598MuP23RnS4+qhcDgc+POf/4zPPvusQ2pAKAxjRKHPLVa8xtYpIca4O94YFQ0DzWCkZFvB53GKImI6PmFsoTdZCB1XxZjW7zwBbwaj58OV1Olz4WjcdGykmWc7JnCqvtnYaKC458B1BgmNkRj7nCpyOrwhJ/XXWr21Roy50XIz8kMosb/Xd9Rh1UfatjUJo6XlMp1iTOc2Y5xfhJABvQccbEr+FqLsmNqx4mJbfo8ddkV4atTJOo/lpTbuPiTEkSNxFaFITBQok5dpbDap3R1t0fX6jlShBat8aibE43HcfPPNKCgokILiOUIarm6GoqIi/PKXv8TSpUvbPWQ4a+JQMTlxZlpNItENpKokGNtbpKPTa32YNP26PQePW2734aKLhDq6tmpWxJirpkeYW6I/avqQJ1cJ50rqlN+LxuKmY+4FcbKEkThhfR7Jv3kuMRSJ6UgSNDbWNtGYWIVeZnTpHhmNrlkrEwrnjS30YsSA3rrXtPNJLkYcdgWFPg8WXzZGEHrKiv26c9u4+xA8Tht8HqdgQirQh3tpzO9IhBFj6H5EYkll+mgs2QmZ/+xmTxkujBsPw/o8TtQHw1i6dhf65edh0cptOkO1NdEx2wxctT5bFfgnn3wS69evx/Lly3tslKi5kKHCboiODBnmqneYbfgv22S906ZgzwOXWCp+bN5XL0I9vEuyMeRFE6Gx8aIYs8+a/d1cWDWPNPvbCLN7aaWawcHzTcYQHK/NAlLJIvSa1T4UaILJQPI7MAvF0f54GJIaehL7j3cSyBWkPVhW7McjV5+T0hHcSkeRw+o3OGOcPjTKj2OFzz//HGPGjMH3vve9Hh8iJMgOyD3ccAHA9u3b8ZWvfAU/+clP8Otf/7pDz+W0+ast3ysu9GLPwS/RGIlhbKFXMLbMYDVxcKNhpUY+b/oozJo4VOTheK6L7zuXnFM2sCrKtjKIlNfJlOMygjP5SCHeKIHEFdLdThsK8vNQ3dAojm+83sWXjcHmffUpOS4yItkK2nID5fM4EY7GTLsZ8++htcG/X7ounqsyLhgK8p1Z5+Zy7YAcj8cxbdo0fPrpp/j444/lfJWAzHFJdGjI0AgeRjJif/0JEXbbWXvM/POJcJPlhEZhyMQfJKy66qNkn6ula3fpyCMqoDNaxLozNm0kCaCyYj/mTR8lPAvKcdHnKfRkHFOhtlYMq71OdHSzMT8ejYsLvSIEyxtb+jxOMW6KxkQfKhpTbopyVjTWtomjPqjdC4ddC9P1dmvHpzGg1bxRt2zqTUaez56DX+LjxEKjJiFoy7sX0/d18Fgj6oNheJx2BEIREUo0ht64kHE6OG2KYBF6nLaU35diMnY7k4XUyzbsNamtSv66GiOxjEbL47SjfJw/bQGyFWSIsOWQHlc3RkezDAnp2IYKrKWZCE6bgnyXw7IlBa18aTVv9KJoH1bHKC7UBFrNCA5cSPXf8y/Ej579IC3jzjjmtTxA5s7IS9fuSmEjcvYfb1ZJn6WapFxZh2ZsPh7yy+Rx6sOrmduXZIKR8Wj0nOn7N55XniPpLWkLAbvmSdoVDOjtQnVDIwp9btQHm4SnR4QK6gtGAsPZghYc6UR2zTyuffv2YezYsZg1axaeeOKJ3G5QN4f0uCQAdDzLkGDFNgQg2F3pEI2rli0p7Iq+tQZn4pkl8s3wcXVAVx9G4yK/V8cwA2Dart1qvHH3oZyMFpBs3snbsBBbD9AUHoyfpXqvcDSe05g7Kir7l8aZWI68ng1QUo6RDrxGiwsH8zMiJurG3YdwvDGiOy8gmcdzOZIyYfkuhxDXjcdVMa5uaBSe3vaagPDEd9YdQ0xNet9pggMCdkX7TZt1ny4r9ut+LxzEIuzXrx9+85vfZD6QhCWkx9UDsGTJEtx99914++23MWHChA45h1y6KxtRzHJffJVvFODldT9GpPO4PCa5Lnrd5dAmWKrzMXZPthpXVh0VxiddexSzcTqYkS5yVeHIRh2EwLtSew33N52yh/E9K0+Ov655zHE0RmJwW5BUjB2gzfJ4nx8OYmt1QJeDM34uXS7VCjZFk5Diory5NIpctmwZbr/9drz++uuYOnVqzsfv7pDkDGm4dOAhwy1btsDj8bT7ORhZfgtf2ZZ1Ej6dQeJIx7rjBA4+qVJ+i3IamcgBnNSRblw+zo/Xd9SJQucRA3oniBPpxzShWhla6itlnJQp78T7TmUzFvfHgizCUVzoTXin2oJhbKEXO+uOiftqdR7G74CfOz+WniCRvh8ZYG7EjWFRY3jZzABnC2PomI6ZjRK8DBFmhjRc0nClYPv27Rg/fjwuv/xyPP3001CyaTvbhlixqSpr42WVqzJOsOkMXHPYalYtVcxyQ8Zxtt6I2dgqh5OJnp7tuZndi2zuD8+DkedEmpNGmF0/bWt17lYGii86MnlcPN9ltZBJp2NpBP8NcE3CdLlMo8d14sQJTJo0CUePHsVHH30k5ycLyByXRAqKiorwxz/+Ef/7v//bYSK8HLMmDtWpLqSDVa4q3+VgrThsgg0IUKsLtxhzJQ2uc2fM9XDQoRTomZE8t2M1HuTVmlVy9QZqHZJu7PM4xaRIY1fi9ckj++sU2WlslqvKdkyf5TqBfH/81pBSBWfo0bb8PDRjq50zV+8gFh4peRjZf40WxiQaU8W2gVBUMCf3HDyeovYBqMIwciV3PRRxrZmWb3zhciTYZJnXtNIlVFUVN954I3bt2oVXXnlFGq1WgjRcPQhXX3015s+fj5/97Gf45z//2dGno1NGSDeB0HsaLV4vFksr51AkLqjZABAMx0RdUTAc0yttGFb6VrwNnn/hBrOGhcDSjVVoSg6hhHoDkFRysB5D0MnNxrSNsf7JCG5orcaFPre49hrWJoTfD7NFg8uhCG+J/lXZ5+IJssPxxgj2HPxSGBcjuAdIf9MiQUFSVWVsoVf0NONG08zQhSJx5Dm0hQBvtaLfJiYkpLL1whWk74hthQceeAAvvPAC/vKXv2DcuHFZHk0iE2SosIchFoth5syZ2LhxI959912ceeaZHXYu6QqTgdTQlM/jQDAcQySu5Un215+wDA0SsaIhFMnYmTlX8PCR1bi58DjtaEzIWlmN2wItvUfpSDO0/7iqmoY5eeiOlyYkv3fN21y9tQZF/iSVnWjvPGfWEliFS+n8cs1rrVy5EjNnzsQvf/nLDmX1dhXIHJc0XGlx7NgxTJw4EbFYDJs2bULfvtYrxrbEiF+sSVEZp+Q5kGTqWRsnazKGx2nHtNEDhTCs+Ta51e4ALZN2yibn1BrGL9tjEYiaHorEdLkpTrYw+5xRoikTiB1qNDT8e+R5Pqqt4jVi3JjxfJ9Vrq254NdLNVvpShmMIcKPP/4Y5513HqZPn44XXngBNpsMbmWCNFzScGXEp59+itLSUpSWlmL16tUd0gfIjCLPC0+HFPSyXEVz78sMfIWfjTfBt0lHVODGzooo0lrGp71g1Ba0IpgYoS8YThoXs/ttU7R9GwuUjdR4h01BNK4m5KC+TKHG0znxpo7FhV5sqwm0yj3n9H8ikiy+bExWLUoIhw8fRmlpKXr37o1///vfOOmkk1p+Yj0AkpwhkRFnnHEGXnjhBbzxxhv4+c9/3iHnQGrh86aPEmOXwyaKQ9OFfo43RkWewgzRuMryMObbcWV6h938UWA8jgQ6lo3ZFuAdiqmBIxFAeKGw8V7YFAVDCnqBaA5Ffo2Iwe8lfSauJo0gb8mS57CLfVAuUYVWJBwSodHk90efi8bioqnjx9WtY7SQOL6CZFF7WbE/J6MViURwxRVX4Pjx41i5cqU0Wm0E6XH1cDzyyCP48Y9/jD//+c+4/vrrO/p08KNnP0iIwKYXOi30uQUBgoNW9kV+b0rvr2ygADqx3+bQ6Ls6rIR+rRpoGun8ZrJS+nqt9CFa+g7SeVxtjWzEcs0wZ84cPPnkk1i3bh0mT57chmfY/SA9Loms8cMf/hA333wzbr31VmzatKmjT0eIwDZGYjh4LGy5nZnRApIr++YYLUALVVHPLruib2xoRDbyQAStkaIj6zE/BnktVmNNwNeW+K91Qr76XJdTUPN9HqeunIB6dNFYgb6Pl9fjgAI9KzOdp0yGbn/9CTRFNY+LSzrxjzbX93XaFFFGQQxGAKInGF2rlXSTFZ588kk89thjeOyxx6TRamM4Mm8i0Z2hKAoee+wx7Ny5E9/61rfw/vvvo7CwsMPOh/TwjBR0IxyGnJSSyCvRar65i/JIXEXJ0L6oC4RQVqypX1idR66xiuON0ZzGgLnSuXEMAAX5LsFwI6kpLs6brQCvGQnBGCrjvaxIC5DGKjSSR3WD5iGZ9dDiZRA2g0Gid3g4cfLI/gD0LVX4tulgRjCJxFVxAyNxVRBuojEVLkeyBU4u2LhxI+bMmYPbb78dt956a06flcgd0uOSgMvlwksvvQSHw4EZM2bg2DHz9iLtg9RJzQzGvBVNfFZFrNnCaVN0rev564phO/LGaAVPhcR8TJ+joltegJtuXMhyLNl8jgu+UjuUEQNOQm+3U9DJzcZDCnohrmpsyXA0hmA4is376rF07S4xPv/B9VixqQorNlXh/AfXo2Ro3xQvzDimwuNCVpRNXo7f507Wbqn6BQAvHqZteIFxrs0k+fb8F8OZoTSOJMScs+lezLF792585zvfwaRJk/Dwww/n9FmJ5kEaLgkAwMCBA7Fq1Srs3bsXl1xyCb788ssOOQ8StfV5nOiT6AllBqsiWTcLlRk9FJpUZ4zzp+0RFkjkdDTBV80Q5rscgqYPaJMceWKRuFZk7HJozDo+jsRVYYA37j4k+lxlGvfLd6Wc1+eHg5af494RKTqQCnq+y4HKqqOm4+01AV2hdEMogtVbk33MqBX90rW7sGjlNlQ3hPD6jgM43hjBkIJeKSLD4WgM4WgcO2uPIa5CF+6lvFZNQ6NQ0SgfpylqKNAMGxk/7btsn+mJk048TntO4cHPPvsMF154Ifr3748XX3wRTqf1b1ai9SDJGRI6vPvuu5g6dSrOPfdcrF69ut0FeY1ivMs27NUpsvN+W2bg4q5G0de6QKMIiWUjsspp7eQ5WJFFFJgL7pohF9V2K10/o1Aw1TbNmz4KS9fuApDsNcWLdo0q9v/8uBaRuCpCrE67gvw8OxpCUV0INhNJJRsSC9+Gyh6MvbaMVHnt+0pf+tBS8HuZCxnjiy++wOTJk+FwOLBx40accsopbXJ+PQWyjksarhbhX//6F775zW9i8uTJeOWVV+Bypa7+2xPcmBlbrmcLPmmmq+tqiYJENgrrQG7KGz6PA8cao1o4j+3TLHdDDTeNdVXcsAEQtXPUCJEjF9Ye39ZpV+Cw2RCKxOC0K4jF1awMnhFtxeKk/YoQJZLF7tyYZ5vfqq2txeTJkxGNRrFx40YMHjy4Dc66Z0GyCiVahK997WuoqKjA//3f/+HKK69EJJJ7C4jWxLINe1HdEMKyDXtNNe8ygQpbCekMk9V7Po9DMP6sgoz8s+mOwd+yGhMaQlHxOt+nWe4mElcFuSUUSZIb6NLD0bh4PxyNo2Ro30SIzi48yiK/V1wnD6eaieny843FVUFZd9iUjHVVVvewrYxW+TgtBOn1OHXjedNHoXRYAfJdjqyN1sGDB/GNb3wDjY2NWL9+vTRaHQBpuCRM8Y1vfAMvv/wy1qxZg6uvvrpDjZexC3GuUAE4WCLD04zcSSAUzchm45NxJqo8z6tYjQE9PZ6ft5E2T8c3k8AiAxaKxITwbCgSQ8VHNVChFfNSCHTPwePiOrmh5Kw/IwOQv6/tOxkiHVvoTYoks5vS2gbK40ySOfiYjlVZdVR4o3y8bMNe3cIoEw4dOoRp06bh6NGjWLduHYYNG9bKVyKRDaThkrDExRdfjL///e9YtWoVvvvd7yIctq6rakvwlhHUeTZX8MmUJKUKfW6hm2dmCDiymWitJvKWgNPjm6Jx09fNjm8Fs/PiBsqKldncmqmPqwOi3ctZp/QR9zsbZFMnR5u4HHZBnuFjzZvUFj18AWQ1Tofa2lpccMEFOHDgANatW4eRI0dmfS0SrQuZ45LIiFdffRXf+ta3cMEFF+Cll15Cr169Mn+oDUEah7xOqbloKyUGM+1CmwLBlDSSJFLZeXEdeSHdZ0glnee9zDoIa8d3pFDK+bbGPJvWrj53MWIz8PyS2bFyBZUCVFYdzUkAtzn44osvcOGFFyIUCmH9+vXSaLUBJDlDGq5Wx/r161FeXo7S0lKsWrWqQzXYzJiHRpJBNij0uVEytEC0y2hpWwwOIgPwdiRkJLk6utXY2On34+qA5bizP8DccPLmnC097/tm5iZ+21zs27cPF16oMQ3XrVuH008/vc2P2RMhDZc0XG2Ct956C5dccgnGjh2LNWvWwOvNPuzT1sjU28sKmkSSpppgpg6fziNr7Ym4o9ARavYtPabP48SHiy5qvROywJ49e3DhhRfC7XZj3bp1GDJkSJsfs6dCsgol2gSTJk3CG2+8gR07duCCCy5AVVVVR5+SgFVBcaq6ux6hiFYXxvM8fD5NF0a0KoLOhGxlnABrLUPO7iMYO0Sbjc3QES1YmnNMuly7gmbnOnPB22+/jUmTJuGkk07Cm2++KY1WJ4I0XBI5obS0FBs3bkQgEMC5556Lt976/+3de1BU1+EH8O8uICCyIOKCKMFolGgFdRCJ/TkjJgoY5BEbq4xMohLtgI80pLRQqxhNqm0zNskoPhg0aprGmmiLaIiA4iMaMahRQEnwEaO8FJVFVJ7n9wfZza4sjwWW3Qvfz8yOl7t37557BvfLuefsOSdNXSQATTNbALojymRoWozQ3Ih2bKtpT3Okva09uk+trkGg+OfRgS1td7W2/jDoKo62VlgbPgaDHW0NXh+rI7Zv3w5/f394enri+PHjcHNzM+r7kWHM7381mT0vLy+cPXsWo0ePxosvvojk5GRTF0mznteK4NGaD7i14WM0U0j1FtrD/lvabk1LI/m0W7TqeRSBX8KzteHonWVrZaH5jpV6dKmx1NfX4/e//z2ioqKwYMECZGZmYuDAgUZ7P+oY9nFRh9XV1eHNN9/E5s2bsXTpUmzYsMHs5mpTr++l7lPRXmvLENrTH1nJZVAqrJtN/6SeZbwt7Z1hoyO6o69N/YXegxeL4erQtC6a5c8LodU1CJ3Vp7VHOraXYwdns+ise/fuYc6cOTh69Cg++ugjREdHQybrpiYlGb+Pa9OmTRg6dChsbGzg5+eHnJycFo/19/eHTCZr9ggODtY57vLlywgNDYWDgwPs7Ozg6+uLmzdvdqR41E2srKyQlJSEpKQkbNmyBUFBQaioqDB1sXTk/ngfAk232QSgWTXXUNq35uoahd45C+vbOa6+vTNsdER3/BUqAGQUlKJBNM3dKPDzpMNas6yrq8LQ0Gp6TfMvCRvb5cuX4efnh3PnziEjIwMxMTEMLTNmcHDt2bMHsbGxSExMxLlz5zB27FgEBgaivLxc7/H79u1DSUmJ5pGXlwcLCwvMnj1bc8zVq1cxefJkPP/888jOzsbFixexcuVK2Ni03KFM5iM6OhqZmZn47rvvMHHiROTn55u6SBrR/sPhaGsFWysLzdIbxvo4ksqti664/q74Xpc+LX1h2JjS0tLg5+cHGxsbnD17FlOnTjXq+1HnGXyr0M/PD76+vti4cSMAoLGxEe7u7li2bBni4+PbfP0HH3yAVatWoaSkBHZ2dgCAuXPnwsrKCrt37+7AJTThrULTu379OsLCwnD9+nV8+umnCAkJMXWR9Prkmx/x3sECzWzoEF3f8jFn5jR031S3BQFACIG///3vSEhIQGhoKHbv3g17e/tueW9qzmi3Cmtra5Gbm4tp06b9cgK5HNOmTcPp06fbdY6UlBTMnTtXE1qNjY04ePAgRo4cicDAQCiVSvj5+eG///1vq+epqamBSqXSeZBpPfvsszh16hSmT5+OsLAw/PWvf4U5dqFGvuABp5/Xu1La20CpaGrZO2qtt9Xael1dTfudtKeeUpdBe+i7hUx3wlu1luY7lMuaD7hwc7TRnFt7oIWjrVWzUYJdO8hCDu+f5y60tZLj3fAxuJAYgI8ixuNCYgAuJAZ0W2g9fvwYkZGRiI+Px4oVK7Bv3z6GloQYFFx3795FQ0MDXFxcdPa7uLigtLS0zdfn5OQgLy8Pb7zxhmZfeXk5Hj58iPXr1yMoKAiHDx/GK6+8glmzZuHYsWMtnmvdunVwcHDQPDhDs3no168fPv/8c6xatQorVqxARESEyRalbI2+uer+EOipCS87a0udIDMm7WjX7hNStwK1+9cahP5JbrW717S3G0XzYfO3HzzRnPv2gyfoY9m0YnNNfQPk7ejX6WiYOdlZo6K6FuLn7e4Kqaep19Hav38/9uzZg7Vr10Iu5wBrKWl9ZtEulpKSAi8vL0ycOFGzr7Gx6V55WFgY3nrrLQDAuHHjcOrUKWzZsgVTpkzRe66EhATExsZqflapVAwvMyGXy7F69WqMGTMGr7/+Ory9vTXfizEXkS946Hxwam9rTyf1/leFmhnV1TfZjNW/YyqtXY++9nJ72tDaC3qq1wnTN59gdxJCYMeOHXjrrbfg4OCAr7/+GuPHj+/2clDnGRRczs7OsLCwQFlZmc7+srIyuLq6tvra6upqfPbZZ1izZk2zc1paWmL06NE6+0eNGtXql1utra1NvsAhte7VV1/FuHHjsHDhQkydOhVLlizB+vXrTTrPYVv0BdrTcyOq+8d6s6eH8Tcf1i/DYEdb3H7wGEqFjc6qwqZoad26dQuLFi1Ceno65s+fjw0bNqB///7dXg7qGga1j/v06QMfHx9kZWVp9jU2NiIrKwuTJk1q9bV79+5FTU0NIiMjm53T19cXhYWFOvu///57eHiY5lYCdZ3nnnsO2dnZ+PDDD7F9+3Z4e3sjOzvb1MUyiPZ6TZuzr+JxXaPRRiZKxdODWbR/lgHdOiqwNUIIbN++Hb/61a9w8eJFpKWlYceOHQwtiTP4xm5sbCySk5Oxc+dOXL58GdHR0aiursaCBQsAAK+99hoSEhKavS4lJQXh4eEYMGBAs+fi4uKwZ88eJCcno6ioCBs3bsSBAwcQExPTgUsicyOXy7F8+XJcvHgRQ4YMwdSpU7F06VKz7PvS5+n+sJ8HImrNFGGhd7s30b5uB1srTcvV2DNdtObWrVt4+eWXERUVhVmzZiEvL6/Z90dJmgzu45ozZw7u3LmDVatWobS0FOPGjUN6erpmwMbNmzebdXQWFhbi5MmTOHz4sN5zvvLKK9iyZQvWrVuH5cuXw9PTE1988QUmT57cgUsic6VufW3cuBHx8fE4dOiQ2fV96fP07UOgqRXm49G/1bWgBtj16dKlUsyNrZUFpo92abEOTEW7L6tfv35IS0tjYPUwnPKJTKKoqAgLFy7EiRMnJNH31RH/t/6IZp0wc/rulKG0y669PdjRVqfvyhywL0u6uKwJmb2e0PfVFvVtRW2OtlaaYfba29q32qz0TIhrZSFr89ak9uua1hnT1dLrnj6HhaxpsUt1+ULGuundNnXLShv7snoXtrjI5LRbX1FRUXjnnXcwePBgUxerS6hHJLZ1W7Gl7Y6+rjPnMFWfVEdduHABcXFxyMzMZCtLwrgCMoNLchobG7F582YkJiaiuroay5cvR3x8PD+AqEVXr17FypUr8e9//xsjRozAP//5T/ZlSRhvFZLkyOVyLFmyBFevXkVcXBw2bdqEYcOGYf369Xj06JGpi0dmpLS0FEuXLsXzzz+PY8eOYevWrcjPz2do9SIMLjIrDg4OWLNmDa5evYp58+Zh5cqVGDFiBLZt24a6ujpTF49MqLKyEitXrsTw4cPxr3/9C++++y5++OEHLF682OzWgSPjYnCRWXJxccHGjRtx5coV+Pv743e/+x3GjBmDvXv3muXEvWQ8T548wYYNGzB8+HC8//77WLZsGa5du4Y//elP6Nu3r6mLRybA4CKzpv7r+vz58xg+fDh++9vfwtfXF5mZmaYuGhlZQ0MDduzYgZEjR+KPf/wjfvOb36CoqAjr169n32cvx+AiSRg3bhwOHTqE7OxsWFpaYvr06Zg+fTq+/vprtsB6mPr6enzxxRfw9vbGwoUL4efnh/z8fGzdurXHjDalzmFwkaRMmTIFp0+fxv79+3H79m1MnjwZ48ePx7Zt2yQzhRTpV1painfffRfPPvssXn31VQwaNAg5OTnYu3cvPD09TV08MiMMLpIcmUyG8PBw5OXlIT09HR4eHoiOjsbgwYOxfPlyXL582dRFpHYSQuD48eOYO3cu3N3dsW7dOsyYMQPnzp1DZmYmfH19TV1EMkP8Hhf1CD/++CO2bduG5ORk3LlzB1OnTkVMTAzCwsI44swMqVQqfPLJJ0hKSkJ+fj5GjhyJmJgYvP7663B0dDR18cgE+D0u6nU8PDzw3nvv4aeffsKnn36Kuro6zJ49Gx4eHli9ejVu375t6iISgEuXLiEmJkbTOvb09ERmZiauXLmCN998k6FF7cIWF/VYFy9exObNm7F79248efIE4eHhiI6Ohr+/Pywsms/lR8bx+PFj/O9//0NSUhJOnDgBV1dXLF68GIsWLcKQIUNMXTwyE5zyicFFWlQqFXbv3o2kpCQUFBTA2dkZwcHBCAkJQUBAAOzt7U1dxB6nrKwMaWlpSE1NRUZGBh4/fgx/f3/ExMQgPDyct2+pGQYXg4v0EELgzJkzSE1NxYEDB5CXl4c+ffrgxRdfRGhoKEJCQtgC6CAhBPLz8zV1e+bMGchkMvz6179GaGgowsLCMHLkSFMXk8wYg4vBRe1w7do1HDhwAKmpqTh27BgaGhowfvx4hIaGIjQ0FOPHj4dM1hvXM26furo6HD9+XFOH169fh52dHQIDAxEaGoqXX34ZAwcONHUxSSIYXAwuMtD9+/eRnp6O1NRUfPnll6isrMTgwYMREhKCmTNnYtKkSXBycjJ1MU1KCIFbt27hxIkTOHDggE49qcPe398fNjY2pi4qSRCDi8FFnVBXV4cTJ04gNTVV05IAgKFDh8LHxwcTJkyAj48PfHx8emyYqUMqNzdX51FeXg4AbJlSl2NwMbioiwgh8MMPP+Dbb7/VfHifO3cOVVVVAHpGmLUVUkqlUnNtEyZMgK+vL9zc3ExcauppGFwMLjKixsZGFBUVtRlmI0aMgJubGwYNGoRBgwZptk1xK62qqgolJSUoLi7W+begoEAnpFxcXDQhpX4MHjyYLSoyOgYXg4u6mTrMtFssN27cQHFxMWpra3WO7d+/f7MwU//r4uICGxsbWFpawtLSElZWVpptCwsL1NfXax51dXWa7UePHukE0tMhVV1drVMGOzs7uLm5YcSIETqtKTc3N4YUmQSDi8FFZkIIgfv377cYKNr/PnnypNPvZ29v3ywM9QUkv7tG5saQz3DLbioTUa8kk8ng5OQEJycnjBkzpsXjhBB48OABysvLUVtbq7dl1dDQoGl9Pd0as7GxgaurK/r169eNV0dkGgwuIjMgk8nQv39/LpBI1A6cZJeIiCSFwUVERJLC4CIiIklhcBERkaQwuIiISFIYXEREJCkMLiIikhQGFxERSQqDi4iIJIXBRUREksLgIiIiSelQcG3atAlDhw6FjY0N/Pz8kJOT0+Kx/v7+kMlkzR7BwcGaY+bPn9/s+aCgoI4UjYiIejiDJ9nds2cPYmNjsWXLFvj5+eGDDz5AYGAgCgsLoVQqmx2/b98+nfWIKioqMHbsWMyePVvnuKCgIOzYsUPzs7W1taFFIyKiXsDgFteGDRuwaNEiLFiwAKNHj8aWLVvQt29fbN++Xe/xTk5OcHV11TwyMjLQt2/fZsFlbW2tcxxnySYiIn0MCq7a2lrk5uZi2rRpv5xALse0adNw+vTpdp0jJSUFc+fOhZ2dnc7+7OxsKJVKeHp6Ijo6GhUVFa2ep6amBiqVSudBREQ9n0HBdffuXTQ0NMDFxUVnv4uLC0pLS9t8fU5ODvLy8vDGG2/o7A8KCsKuXbuQlZWFv/3tbzh27BhmzJiBhoaGFs+1bt06ODg4aB7u7u6GXAoREUlUty4kmZKSAi8vL0ycOFFn/9y5czXbXl5e8Pb2xvDhw5GdnY2XXnpJ77kSEhIQGxur+VmlUjG8iIh6AYNaXM7OzrCwsEBZWZnO/rKyMri6urb62urqanz22WeIiopq832GDRsGZ2dnFBUVtXiMtbU1FAqFzoOIiHo+g4KrT58+8PHxQVZWlmZfY2MjsrKyMGnSpFZfu3fvXtTU1CAyMrLN97l16xYqKiowaNAgQ4pHRES9gMGjCmNjY5GcnIydO3fi8uXLiI6ORnV1NRYsWAAAeO2115CQkNDsdSkpKQgPD8eAAQN09j98+BBxcXH45ptvcOPGDWRlZSEsLAzPPfccAgMDO3hZRETUUxncxzVnzhzcuXMHq1atQmlpKcaNG4f09HTNgI2bN29CLtfNw8LCQpw8eRKHDx9udj4LCwtcvHgRO3fuxIMHD+Dm5oaAgACsXbuW3+UiIqJmZEIIYepCdAWVSgUHBwdUVlayv4uISGIM+QznXIVERCQpDC4iIpIUBhcREUkKg4uIiCSFwUVERJLC4CIiIklhcBERkaQwuIiISFIYXEREJCkMLiIikhQGFxERSQqDi4iIJIXBRUREksLgIiIiSWFwERGRpDC4iIhIUhhcREQkKQwuIiKSFAYXERFJCoOLiIgkxdLUBegqQggAgEqlMnFJiIjIUOrPbvVneWt6THBVVVUBANzd3U1cEiIi6qiqqio4ODi0eoxMtCfeJKCxsRHFxcWwt7eHTCYzdXG6nUqlgru7O3766ScoFApTF6fHY313H9Z19zFlXQshUFVVBTc3N8jlrfdi9ZgWl1wux5AhQ0xdDJNTKBT8z92NWN/dh3XdfUxV1221tNQ4OIOIiCSFwUVERJLC4OohrK2tkZiYCGtra1MXpVdgfXcf1nX3kUpd95jBGURE1DuwxUVERJLC4CIiIklhcBERkaQwuIiISFIYXBJy7949zJs3DwqFAo6OjoiKisLDhw9bPX7ZsmXw9PSEra0tnnnmGSxfvhyVlZV6j6+oqMCQIUMgk8nw4MEDI12FNBijrr/77jtERETA3d0dtra2GDVqFD788MPuuByzZqzf65s3byI4OBh9+/aFUqlEXFwc6uvrjX05Zs/Q+gaAbdu2wd/fHwqFosXPh++//x5hYWFwdnaGQqHA5MmTcfToUaNcA4NLQubNm4f8/HxkZGQgLS0Nx48fx+LFi1s8vri4GMXFxXj//feRl5eHjz/+GOnp6YiKitJ7fFRUFLy9vY1VfEkxRl3n5uZCqVTik08+QX5+PlasWIGEhARs3LixOy7JbBmjrhsaGhAcHIza2lqcOnUKO3fuxMcff4xVq1Z1xyWZNUPrGwAePXqEoKAg/PnPf27xmJkzZ6K+vh5HjhxBbm4uxo4di5kzZ6K0tLSrLwEQJAkFBQUCgDh79qxm35dffilkMpm4fft2u8/zn//8R/Tp00fU1dXp7E9KShJTpkwRWVlZAoC4f/9+VxVdcoxd19piYmLE1KlTO1VeKTNWXR86dEjI5XJRWlqqOWbz5s1CoVCImpqarrsAielsfR89elTv58OdO3cEAHH8+HHNPpVKJQCIjIyMLiu/GltcEnH69Gk4OjpiwoQJmn3Tpk2DXC7HmTNn2n2eyspKKBQKWFr+Mk1lQUEB1qxZg127drU5uWVvYMy61neMk5NTp8orZcaq69OnT8PLywsuLi6aYwIDA6FSqZCfn991FyAxXVXfTxswYAA8PT2xa9cuVFdXo76+Hlu3boVSqYSPj09XFF1Hj5lkt6crLS2FUqnU2WdpaQknJ6d2N8Xv3r2LtWvX6twWqKmpQUREBP7xj3/gmWeewbVr17q03FJkrLp+2qlTp7Bnzx4cPHiwU+WVMmPVdWlpqU5oAdD8bJRbVxLRFfWtj0wmQ2ZmJsLDw2Fvbw+5XA6lUon09HT079+/s8Vuhn9em1h8fDxkMlmrjytXrnT6fVQqFYKDgzF69GisXr1asz8hIQGjRo1CZGRkp9/D3Jm6rrXl5eUhLCwMiYmJCAgI6PR7mhtzquveoLvquyVCCCxZsgRKpRInTpxATk4OwsPDERISgpKSki5/P7a4TOztt9/G/PnzWz1m2LBhcHV1RXl5uc7++vp63Lt3D66urq2+vqqqCkFBQbC3t8f+/fthZWWlee7IkSO4dOkSPv/8cwC/rD7q7OyMFStW4J133unAVZknU9e1WkFBAV566SUsXrwYf/nLXwy+DikwdV27uroiJydH5/iysjLNcz1Nd9R3a44cOYK0tDTcv39fsxxKUlISMjIysHPnTsTHx3f43Hp1ea8ZGYW6U/Xbb7/V7Pvqq6/a7FStrKwUL7zwgpgyZYqorq5u9nxRUZG4dOmS5rF9+3YBQJw6dUqUlZUZ5VrMnbHqWggh8vLyhFKpFHFxcV1ebikyVl2rB2do/w5v3bpVKBQK8eTJk669CAnpaH2rtTQ4IzU1VcjlclFVVaWzf+TIkeK9997rkrJrY3BJSFBQkBg/frw4c+aMOHnypBgxYoSIiIjQPH/r1i3h6ekpzpw5I4Ro+s/t5+cnvLy8RFFRkSgpKdE86uvr9b5HS7+YvY0x6vrSpUti4MCBIjIyUuf58vJyk1yjuTBGXdfX14sxY8aIgIAAceHCBZGeni4GDhwoEhISTHKN5sTQ+hZCiJKSEnH+/HmRnJysGT14/vx5UVFRIYRoGlU4YMAAMWvWLHHhwgVRWFgo/vCHPwgrKytx4cKFLr8GBpeEVFRUiIiICNGvXz+hUCjEggULdP7CuX79ugAgjh49KoT4JYT0Pa5fv673PRhcTYxR14mJiXqf9/Dw6P4LNCPG+r2+ceOGmDFjhrC1tRXOzs7i7bffbvWrCb2FofUtRMu/uzt27NAcc/bsWREQECCcnJyEvb29eOGFF8ShQ4eMcg1c1oSIiCSFowqJiEhSGFxERCQpDC4iIpIUBhcREUkKg4uIiCSFwUVERJLC4CIiIklhcBERkaQwuIiISFIYXEREJCkMLiIikhQGFxERScr/AzX+72/l2vmQAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def project_XY(cleaned_point_cloud, VIZ=False):\n",
    "    # assume cleaned_point_cloud is the cleaned point cloud as a numpy array with shape (n_points, 3)\n",
    "\n",
    "    # project points onto the XY plane by ignoring the Z coordinate\n",
    "    projected_points = cleaned_point_cloud[:, [1,0]]\n",
    "\n",
    "    # compute extent of points in XY plane\n",
    "    min_x, min_y = np.min(projected_points, axis=0)\n",
    "    max_x, max_y = np.max(projected_points, axis=0)\n",
    "\n",
    "    # compute center of circle\n",
    "    center_x, center_y = (min_x + max_x) / 2, (min_y + max_y) / 2\n",
    "\n",
    "    # compute radius of circle\n",
    "    radius = (max_x - min_x)/ 2\n",
    "    # plot the circle and points\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.set_aspect('equal')\n",
    "    ax.scatter(projected_points[:, 0], projected_points[:, 1], s=1)\n",
    "    circle = plt.Circle((center_x, center_y), radius, fill=False)\n",
    "    ax.add_artist(circle)\n",
    "\n",
    "    edge = 0.01\n",
    "    plt.xlim(center_x-radius-edge, center_x+radius+edge)\n",
    "    plt.ylim(center_y-radius-edge, center_y+radius+edge)\n",
    "    # plt.axis('equal')\n",
    "    plt.show()\n",
    "    return center_x, center_y, radius\n",
    "\n",
    "center_x, center_y, radius = project_XY(cleaned_point_cloud, VIZ=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def get_ceneter_position(cleaned_point_cloud, VIZ=False):\n",
    "    center_z = project_YZ(cleaned_point_cloud, VIZ=VIZ)\n",
    "    center_x, center_y, radius = project_XY(cleaned_point_cloud, VIZ=VIZ)\n",
    "    center_position_array = np.array([center_y,center_x, center_z])\n",
    "\n",
    "    return center_position_array, radius\n",
    "\n",
    "center_position_array, radius = get_ceneter_position(cleaned_point_cloud, VIZ=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save(\"center_position_array.npy\", center_position_array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_camera(perception_path, camera_p, rotation_mat):\n",
    "    CAPTURE = True\n",
    "\n",
    "    import subprocess\n",
    "    import os\n",
    "\n",
    "    if CAPTURE:\n",
    "        if os.path.exists(f\"{perception_path}/label_data.npy\"):\n",
    "            os.remove(f\"{perception_path}/label_data.npy\")\n",
    "\n",
    "        if os.path.exists(f\"{perception_path}/depth_data.npy\"):\n",
    "            os.remove(f\"{perception_path}/depth_data.npy\")\n",
    "\n",
    "        subprocess.run([f\"cd {perception_path}; ./experiments/scripts/ros_seg_rgbd_add_test_segmentation_realsense.sh $GPU_ID 0\"], shell=True)\n",
    "    \n",
    "    label_pixel = np.load(f'{perception}/label_data.npy')\n",
    "    depth_pixel = np.load(f'{perception}/depth_data.npy') \n",
    "\n",
    "    position_mat = Translation(*(camera_p.tolist()))\n",
    "    transform_mat= HT_matrix(rotation_mat, position_mat)\n",
    "\n",
    "        \n",
    "    clustering_num=len(np.unique(label_pixel));target_cluster_center_point=None;cluster_cen_pose_lst = []\n",
    "    # Total cluster loop\n",
    "    for cluster_idx in range(clustering_num):\n",
    "        if cluster_idx ==0: continue\n",
    "\n",
    "        # Random color \n",
    "        # color = random_color_gen()\n",
    "        cluster_xlst, cluster_ylst = np.where(label_pixel==cluster_idx)\n",
    "\n",
    "        clusters = []\n",
    "        for idx, (x,y,z) in enumerate(zip(depth_pixel[2,cluster_xlst, cluster_ylst], \\\n",
    "                                            depth_pixel[0,cluster_xlst, cluster_ylst], \\\n",
    "                                            depth_pixel[1,cluster_xlst, cluster_ylst])):\n",
    "            position=camera_to_base(transform_mat, np.array([[x,y,z]])).reshape(-1)\n",
    "            clusters.append(position)\n",
    "\n",
    "        break\n",
    "            \n",
    "    single_cluster = np.stack(clusters)\n",
    "    cleaned_point_cloud = clean_point_cloud(single_cluster)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}