main.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e962e9e2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda\n",
      "GeForce RTX 2080 Ti\n"
     ]
    }
   ],
   "source": [
    "## Training Script\n",
    "# Run this to train a bone suppression network ResNet-BS, introduced by Rajaraman et al. 2021.\n",
    "# \n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from pathlib import Path\n",
    "import os, sys, datetime, time, random, fnmatch, math\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.animation as animation\n",
    "import skimage.metrics\n",
    "from sklearn.model_selection import KFold\n",
    "\n",
    "import torch\n",
    "from torchvision import transforms as tvtransforms\n",
    "from torch.utils.data import Dataset, DataLoader, random_split, SubsetRandomSampler, ConcatDataset\n",
    "import torchvision.utils as vutils\n",
    "import torch.utils.tensorboard as tensorboard\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "\n",
    "import datasets, custom_transforms, RajaramanModel, pytorch_msssim\n",
    "\n",
    "flag_debug = False\n",
    "flag_load_previous_save = False\n",
    "\n",
    "# Input Directories\n",
    "#data_BSE = \"D:/data/JSRT/augmented/train/target/\"\n",
    "#data_normal = \"D:/data/JSRT/augmented/train/source/\"\n",
    "data_BSE = Path('G:/DanielLam/JSRT/HQ_JSRT_and_BSE-JSRT/177-20-20 split/trainValidate_LowRes/suppressed/')\n",
    "data_normal = Path('G:/DanielLam/JSRT/HQ_JSRT_and_BSE-JSRT/177-20-20 split/trainValidate_LowRes/normal/')\n",
    "#data_val_normal = 'G:/DanielLam/JSRT/HQ_JSRT_and_BSE-JSRT/177-20-20 split/validate/normal'\n",
    "#data_val_BSE = 'G:/DanielLam/JSRT/HQ_JSRT_and_BSE-JSRT/177-20-20 split/validate/suppressed/'\n",
    "\n",
    "# Save directories:\n",
    "output_save_directory = Path(\"./runs/Rajaraman_ResNet/177-20-20/10KFold_217perEpoch\")\n",
    "output_save_directory.mkdir(parents=True, exist_ok=True)\n",
    "PATH_SAVE_NETWORK_INTERMEDIATE = os.path.join(output_save_directory, 'network_intermediate_{}.tar' )\n",
    "PATH_SAVE_NETWORK_FINAL = os.path.join(output_save_directory, 'network_final_{}.pt')\n",
    "\n",
    "# Image Size:\n",
    "image_spatial_size = (480, 480)\n",
    "_batch_size = 2\n",
    "split_k_folds=10\n",
    "sample_keys_images = [\"source\", \"boneless\"]\n",
    "\n",
    "# Optimisation\n",
    "lr_ini = 0.001\n",
    "beta1 = 0.9\n",
    "beta2 = 0.999\n",
    "\n",
    "# Training\n",
    "total_num_epochs_paper = 200\n",
    "epoch_factor = 1 #22.8764\n",
    "total_num_epochs = int(epoch_factor*total_num_epochs_paper)\n",
    "\n",
    "# Weight Initialisation\n",
    "def weights_init(m):\n",
    "    classname = m.__class__.__name__\n",
    "    if isinstance(m, nn.Conv2d) or isinstance(m, nn.ConvTranspose2d):\n",
    "        nn.init.normal_(m.weight.data, 0., 0.02)\n",
    "        #nn.init.kaiming_normal_(m.weight.data,0)\n",
    "        try:\n",
    "            nn.init.constant_(m.bias.data, 0.)\n",
    "        except:\n",
    "            pass\n",
    "    if isinstance(m, nn.BatchNorm2d) or isinstance(m, nn.InstanceNorm2d):\n",
    "        if m.affine:\n",
    "            nn.init.normal_(m.weight.data, 1.0, 0.02)\n",
    "            nn.init.constant_(m.bias.data, 0.)\n",
    "\n",
    "## Code for putting things on the GPU\n",
    "ngpu = 1 #torch.cuda.device_count()\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0,1\"\n",
    "device = torch.device(\"cuda\" if (torch.cuda.is_available() and ngpu > 0) else \"cpu\")\n",
    "print(device)\n",
    "if (torch.cuda.is_available()):\n",
    "    print(torch.cuda.get_device_name(torch.cuda.current_device()))\n",
    "    if ngpu ==1:\n",
    "        device=torch.device('cuda:1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7e173907",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Current date:\n",
    "current_date=datetime.datetime.today().strftime('%Y-%m-%d')\n",
    "\n",
    "# Data Loader\n",
    "original_key = \"source\"\n",
    "target_key = \"boneless\"\n",
    "discriminator_keys_images = [original_key, target_key]\n",
    "\n",
    "#ds_training = datasets.JSRT_CXR(data_normal, data_BSE,\n",
    "#                         transform=tvtransforms.Compose([\n",
    "#                             # custom_transforms.HistogramEqualisation(discriminator_keys_images),-- check if training data is already equalised\n",
    "#                             custom_transforms.Resize(discriminator_keys_images, image_spatial_size),\n",
    "#                             custom_transforms.ToTensor(discriminator_keys_images),\n",
    "#                             ])\n",
    "#                      )\n",
    "#ds_val = datasets.JSRT_CXR(data_val_normal, data_val_BSE,\n",
    "#                         transform=tvtransforms.Compose([\n",
    "#                             #custom_transforms.HistogramEqualisation(discriminator_keys_images),\n",
    "#                             custom_transforms.Resize(discriminator_keys_images, image_spatial_size),\n",
    "#                             custom_transforms.ToTensor(discriminator_keys_images),\n",
    "#                             ])\n",
    "#                      )\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "9bf0b2c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rajaraman Loss\n",
    "def criterion_Rajaraman(testImage, referenceImage, alpha=0.84):\n",
    "    \"\"\"\n",
    "    Generates a loss over a mini-batch of images.\n",
    "    Outputs scalars.\n",
    "    Rajaraman, S., Zamzmi, G., Folio, L., Alderson, P., & Antani, S. (2021). Chest X-Ray Bone Suppression for Improving Classification of Tuberculosis-Consistent Findings. In Diagnostics (Vol. 11, Issue 5). https://doi.org/10.3390/diagnostics11050840\n",
    "    \"\"\"\n",
    "    mae = nn.L1Loss() # L2 used for easier optimisation c.f. L1\n",
    "    mae_loss = mae(testImage, referenceImage)\n",
    "    msssim = pytorch_msssim.MSSSIM(window_size=11, size_average=True, channel=1, normalize='relu')\n",
    "    msssim_loss = 1 - msssim(testImage, referenceImage)\n",
    "    total_loss = (1-alpha)*mae_loss + alpha*msssim_loss\n",
    "    return total_loss, mae_loss, msssim_loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "72a01323",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "boneless\n",
      "Fold 1\n",
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.7445\tMAELoss: 0.3417\tMSSSIM Loss: 0.8212\n",
      "[0/200][1/89]\tTotal Loss: 0.5659\tMAELoss: 0.3097\tMSSSIM Loss: 0.6147\n",
      "[0/200][2/89]\tTotal Loss: 0.5465\tMAELoss: 0.2771\tMSSSIM Loss: 0.5978\n",
      "[0/200][3/89]\tTotal Loss: 0.4427\tMAELoss: 0.2620\tMSSSIM Loss: 0.4772\n",
      "[0/200][4/89]\tTotal Loss: 0.3007\tMAELoss: 0.1678\tMSSSIM Loss: 0.3260\n",
      "[0/200][5/89]\tTotal Loss: 0.3490\tMAELoss: 0.2718\tMSSSIM Loss: 0.3637\n",
      "[0/200][6/89]\tTotal Loss: 0.3133\tMAELoss: 0.1506\tMSSSIM Loss: 0.3443\n",
      "[0/200][7/89]\tTotal Loss: 0.2097\tMAELoss: 0.2536\tMSSSIM Loss: 0.2014\n",
      "[0/200][8/89]\tTotal Loss: 0.2371\tMAELoss: 0.1905\tMSSSIM Loss: 0.2460\n",
      "[0/200][9/89]\tTotal Loss: 0.1794\tMAELoss: 0.2561\tMSSSIM Loss: 0.1648\n",
      "[0/200][10/89]\tTotal Loss: 0.1442\tMAELoss: 0.1284\tMSSSIM Loss: 0.1472\n",
      "[0/200][11/89]\tTotal Loss: 0.1196\tMAELoss: 0.0938\tMSSSIM Loss: 0.1245\n",
      "[0/200][12/89]\tTotal Loss: 0.1141\tMAELoss: 0.0762\tMSSSIM Loss: 0.1213\n",
      "[0/200][13/89]\tTotal Loss: 0.2345\tMAELoss: 0.1609\tMSSSIM Loss: 0.2485\n",
      "[0/200][14/89]\tTotal Loss: 0.0840\tMAELoss: 0.0853\tMSSSIM Loss: 0.0837\n",
      "[0/200][15/89]\tTotal Loss: 0.0707\tMAELoss: 0.0587\tMSSSIM Loss: 0.0730\n",
      "[0/200][16/89]\tTotal Loss: 0.0399\tMAELoss: 0.0690\tMSSSIM Loss: 0.0344\n",
      "[0/200][17/89]\tTotal Loss: 0.0676\tMAELoss: 0.0784\tMSSSIM Loss: 0.0655\n",
      "[0/200][18/89]\tTotal Loss: 0.0672\tMAELoss: 0.0710\tMSSSIM Loss: 0.0665\n",
      "[0/200][19/89]\tTotal Loss: 0.0732\tMAELoss: 0.0818\tMSSSIM Loss: 0.0715\n",
      "[0/200][20/89]\tTotal Loss: 0.0898\tMAELoss: 0.1117\tMSSSIM Loss: 0.0857\n",
      "[0/200][21/89]\tTotal Loss: 0.0898\tMAELoss: 0.0994\tMSSSIM Loss: 0.0880\n",
      "[0/200][22/89]\tTotal Loss: 0.0529\tMAELoss: 0.0817\tMSSSIM Loss: 0.0474\n",
      "[0/200][23/89]\tTotal Loss: 0.0452\tMAELoss: 0.0508\tMSSSIM Loss: 0.0441\n",
      "[0/200][24/89]\tTotal Loss: 0.0410\tMAELoss: 0.0433\tMSSSIM Loss: 0.0405\n",
      "[0/200][25/89]\tTotal Loss: 0.0407\tMAELoss: 0.0772\tMSSSIM Loss: 0.0338\n",
      "[0/200][26/89]\tTotal Loss: 0.0417\tMAELoss: 0.0701\tMSSSIM Loss: 0.0363\n",
      "[0/200][27/89]\tTotal Loss: 0.0330\tMAELoss: 0.0503\tMSSSIM Loss: 0.0297\n",
      "[0/200][28/89]\tTotal Loss: 0.0561\tMAELoss: 0.0542\tMSSSIM Loss: 0.0565\n",
      "[0/200][29/89]\tTotal Loss: 0.0493\tMAELoss: 0.0415\tMSSSIM Loss: 0.0508\n",
      "[0/200][30/89]\tTotal Loss: 0.0363\tMAELoss: 0.0312\tMSSSIM Loss: 0.0373\n",
      "[0/200][31/89]\tTotal Loss: 0.0369\tMAELoss: 0.0451\tMSSSIM Loss: 0.0354\n",
      "[0/200][32/89]\tTotal Loss: 0.0478\tMAELoss: 0.0628\tMSSSIM Loss: 0.0450\n",
      "[0/200][33/89]\tTotal Loss: 0.0317\tMAELoss: 0.0345\tMSSSIM Loss: 0.0312\n",
      "[0/200][34/89]\tTotal Loss: 0.0437\tMAELoss: 0.0556\tMSSSIM Loss: 0.0415\n",
      "[0/200][35/89]\tTotal Loss: 0.0414\tMAELoss: 0.0487\tMSSSIM Loss: 0.0400\n",
      "[0/200][36/89]\tTotal Loss: 0.0516\tMAELoss: 0.0574\tMSSSIM Loss: 0.0505\n",
      "[0/200][37/89]\tTotal Loss: 0.0374\tMAELoss: 0.0523\tMSSSIM Loss: 0.0346\n",
      "[0/200][38/89]\tTotal Loss: 0.0379\tMAELoss: 0.0375\tMSSSIM Loss: 0.0380\n",
      "[0/200][39/89]\tTotal Loss: 0.0321\tMAELoss: 0.0454\tMSSSIM Loss: 0.0296\n",
      "[0/200][40/89]\tTotal Loss: 0.0420\tMAELoss: 0.0722\tMSSSIM Loss: 0.0362\n",
      "[0/200][41/89]\tTotal Loss: 0.0424\tMAELoss: 0.0484\tMSSSIM Loss: 0.0413\n",
      "[0/200][42/89]\tTotal Loss: 0.0330\tMAELoss: 0.0386\tMSSSIM Loss: 0.0319\n",
      "[0/200][43/89]\tTotal Loss: 0.0547\tMAELoss: 0.0415\tMSSSIM Loss: 0.0572\n",
      "[0/200][44/89]\tTotal Loss: 0.0512\tMAELoss: 0.0370\tMSSSIM Loss: 0.0539\n",
      "[0/200][45/89]\tTotal Loss: 0.0461\tMAELoss: 0.0511\tMSSSIM Loss: 0.0451\n",
      "[0/200][46/89]\tTotal Loss: 0.0426\tMAELoss: 0.0613\tMSSSIM Loss: 0.0391\n",
      "[0/200][47/89]\tTotal Loss: 0.0491\tMAELoss: 0.0784\tMSSSIM Loss: 0.0435\n",
      "[0/200][48/89]\tTotal Loss: 0.0321\tMAELoss: 0.0423\tMSSSIM Loss: 0.0302\n",
      "[0/200][49/89]\tTotal Loss: 0.0327\tMAELoss: 0.0361\tMSSSIM Loss: 0.0321\n",
      "[0/200][50/89]\tTotal Loss: 0.0481\tMAELoss: 0.0451\tMSSSIM Loss: 0.0486\n",
      "[0/200][51/89]\tTotal Loss: 0.0355\tMAELoss: 0.0415\tMSSSIM Loss: 0.0344\n",
      "[0/200][52/89]\tTotal Loss: 0.0346\tMAELoss: 0.0349\tMSSSIM Loss: 0.0346\n",
      "[0/200][53/89]\tTotal Loss: 0.0333\tMAELoss: 0.0468\tMSSSIM Loss: 0.0307\n",
      "[0/200][54/89]\tTotal Loss: 0.0411\tMAELoss: 0.0347\tMSSSIM Loss: 0.0423\n",
      "[0/200][55/89]\tTotal Loss: 0.0372\tMAELoss: 0.0535\tMSSSIM Loss: 0.0341\n",
      "[0/200][56/89]\tTotal Loss: 0.0437\tMAELoss: 0.0629\tMSSSIM Loss: 0.0400\n",
      "[0/200][57/89]\tTotal Loss: 0.0234\tMAELoss: 0.0361\tMSSSIM Loss: 0.0210\n",
      "[0/200][58/89]\tTotal Loss: 0.0342\tMAELoss: 0.0271\tMSSSIM Loss: 0.0355\n",
      "[0/200][59/89]\tTotal Loss: 0.0348\tMAELoss: 0.0352\tMSSSIM Loss: 0.0347\n",
      "[0/200][60/89]\tTotal Loss: 0.0343\tMAELoss: 0.0283\tMSSSIM Loss: 0.0354\n",
      "[0/200][61/89]\tTotal Loss: 0.0471\tMAELoss: 0.0467\tMSSSIM Loss: 0.0472\n",
      "[0/200][62/89]\tTotal Loss: 0.0320\tMAELoss: 0.0516\tMSSSIM Loss: 0.0282\n",
      "[0/200][63/89]\tTotal Loss: 0.0302\tMAELoss: 0.0376\tMSSSIM Loss: 0.0288\n",
      "[0/200][64/89]\tTotal Loss: 0.0223\tMAELoss: 0.0269\tMSSSIM Loss: 0.0214\n",
      "[0/200][65/89]\tTotal Loss: 0.0386\tMAELoss: 0.0314\tMSSSIM Loss: 0.0399\n",
      "[0/200][66/89]\tTotal Loss: 0.0232\tMAELoss: 0.0222\tMSSSIM Loss: 0.0234\n",
      "[0/200][67/89]\tTotal Loss: 0.0372\tMAELoss: 0.0355\tMSSSIM Loss: 0.0376\n",
      "[0/200][68/89]\tTotal Loss: 0.0231\tMAELoss: 0.0351\tMSSSIM Loss: 0.0209\n",
      "[0/200][69/89]\tTotal Loss: 0.0168\tMAELoss: 0.0284\tMSSSIM Loss: 0.0146\n",
      "[0/200][70/89]\tTotal Loss: 0.0237\tMAELoss: 0.0228\tMSSSIM Loss: 0.0239\n",
      "[0/200][71/89]\tTotal Loss: 0.0351\tMAELoss: 0.0244\tMSSSIM Loss: 0.0371\n",
      "[0/200][72/89]\tTotal Loss: 0.0260\tMAELoss: 0.0315\tMSSSIM Loss: 0.0250\n",
      "[0/200][73/89]\tTotal Loss: 0.0329\tMAELoss: 0.0324\tMSSSIM Loss: 0.0330\n",
      "[0/200][74/89]\tTotal Loss: 0.0283\tMAELoss: 0.0372\tMSSSIM Loss: 0.0266\n",
      "[0/200][75/89]\tTotal Loss: 0.0371\tMAELoss: 0.0446\tMSSSIM Loss: 0.0356\n",
      "[0/200][76/89]\tTotal Loss: 0.0367\tMAELoss: 0.0416\tMSSSIM Loss: 0.0358\n",
      "[0/200][77/89]\tTotal Loss: 0.0168\tMAELoss: 0.0155\tMSSSIM Loss: 0.0170\n",
      "[0/200][78/89]\tTotal Loss: 0.0370\tMAELoss: 0.0361\tMSSSIM Loss: 0.0372\n",
      "[0/200][79/89]\tTotal Loss: 0.0296\tMAELoss: 0.0296\tMSSSIM Loss: 0.0296\n",
      "[0/200][80/89]\tTotal Loss: 0.0372\tMAELoss: 0.0368\tMSSSIM Loss: 0.0373\n",
      "[0/200][81/89]\tTotal Loss: 0.0380\tMAELoss: 0.0447\tMSSSIM Loss: 0.0367\n",
      "[0/200][82/89]\tTotal Loss: 0.0306\tMAELoss: 0.0392\tMSSSIM Loss: 0.0290\n",
      "[0/200][83/89]\tTotal Loss: 0.0335\tMAELoss: 0.0364\tMSSSIM Loss: 0.0330\n",
      "[0/200][84/89]\tTotal Loss: 0.0239\tMAELoss: 0.0230\tMSSSIM Loss: 0.0241\n",
      "[0/200][85/89]\tTotal Loss: 0.0344\tMAELoss: 0.0288\tMSSSIM Loss: 0.0355\n",
      "[0/200][86/89]\tTotal Loss: 0.0305\tMAELoss: 0.0344\tMSSSIM Loss: 0.0297\n",
      "[0/200][87/89]\tTotal Loss: 0.0245\tMAELoss: 0.0288\tMSSSIM Loss: 0.0236\n",
      "[0/200][88/89]\tTotal Loss: 0.0212\tMAELoss: 0.0332\tMSSSIM Loss: 0.0189\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    28: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    58: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    69: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    80: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    91: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch   106: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   117: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_1.tar\n",
      "Fold 2\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.7162\tMAELoss: 0.3250\tMSSSIM Loss: 0.7907\n",
      "[0/200][1/89]\tTotal Loss: 0.6091\tMAELoss: 0.2805\tMSSSIM Loss: 0.6717\n",
      "[0/200][2/89]\tTotal Loss: 0.5446\tMAELoss: 0.2719\tMSSSIM Loss: 0.5966\n",
      "[0/200][3/89]\tTotal Loss: 0.3410\tMAELoss: 0.1624\tMSSSIM Loss: 0.3750\n",
      "[0/200][4/89]\tTotal Loss: 0.2570\tMAELoss: 0.2266\tMSSSIM Loss: 0.2628\n",
      "[0/200][5/89]\tTotal Loss: 0.3211\tMAELoss: 0.3008\tMSSSIM Loss: 0.3249\n",
      "[0/200][6/89]\tTotal Loss: 0.1586\tMAELoss: 0.3200\tMSSSIM Loss: 0.1279\n",
      "[0/200][7/89]\tTotal Loss: 0.2148\tMAELoss: 0.1757\tMSSSIM Loss: 0.2222\n",
      "[0/200][8/89]\tTotal Loss: 0.1752\tMAELoss: 0.2093\tMSSSIM Loss: 0.1687\n",
      "[0/200][9/89]\tTotal Loss: 0.1298\tMAELoss: 0.1364\tMSSSIM Loss: 0.1285\n",
      "[0/200][10/89]\tTotal Loss: 0.1296\tMAELoss: 0.0796\tMSSSIM Loss: 0.1391\n",
      "[0/200][11/89]\tTotal Loss: 0.1311\tMAELoss: 0.0958\tMSSSIM Loss: 0.1378\n",
      "[0/200][12/89]\tTotal Loss: 0.0849\tMAELoss: 0.0859\tMSSSIM Loss: 0.0847\n",
      "[0/200][13/89]\tTotal Loss: 0.0591\tMAELoss: 0.0850\tMSSSIM Loss: 0.0542\n",
      "[0/200][14/89]\tTotal Loss: 0.0985\tMAELoss: 0.1027\tMSSSIM Loss: 0.0977\n",
      "[0/200][15/89]\tTotal Loss: 0.0636\tMAELoss: 0.0758\tMSSSIM Loss: 0.0612\n",
      "[0/200][16/89]\tTotal Loss: 0.0565\tMAELoss: 0.0592\tMSSSIM Loss: 0.0560\n",
      "[0/200][17/89]\tTotal Loss: 0.0822\tMAELoss: 0.0831\tMSSSIM Loss: 0.0820\n",
      "[0/200][18/89]\tTotal Loss: 0.0476\tMAELoss: 0.0538\tMSSSIM Loss: 0.0464\n",
      "[0/200][19/89]\tTotal Loss: 0.0419\tMAELoss: 0.0388\tMSSSIM Loss: 0.0424\n",
      "[0/200][20/89]\tTotal Loss: 0.0443\tMAELoss: 0.0484\tMSSSIM Loss: 0.0436\n",
      "[0/200][21/89]\tTotal Loss: 0.0569\tMAELoss: 0.0882\tMSSSIM Loss: 0.0509\n",
      "[0/200][22/89]\tTotal Loss: 0.0420\tMAELoss: 0.0659\tMSSSIM Loss: 0.0374\n",
      "[0/200][23/89]\tTotal Loss: 0.0495\tMAELoss: 0.0585\tMSSSIM Loss: 0.0478\n",
      "[0/200][24/89]\tTotal Loss: 0.0345\tMAELoss: 0.0276\tMSSSIM Loss: 0.0359\n",
      "[0/200][25/89]\tTotal Loss: 0.0272\tMAELoss: 0.0304\tMSSSIM Loss: 0.0266\n",
      "[0/200][26/89]\tTotal Loss: 0.0995\tMAELoss: 0.0891\tMSSSIM Loss: 0.1015\n",
      "[0/200][27/89]\tTotal Loss: 0.0333\tMAELoss: 0.0309\tMSSSIM Loss: 0.0338\n",
      "[0/200][28/89]\tTotal Loss: 0.0408\tMAELoss: 0.0402\tMSSSIM Loss: 0.0409\n",
      "[0/200][29/89]\tTotal Loss: 0.0450\tMAELoss: 0.0705\tMSSSIM Loss: 0.0402\n",
      "[0/200][30/89]\tTotal Loss: 0.0551\tMAELoss: 0.0759\tMSSSIM Loss: 0.0511\n",
      "[0/200][31/89]\tTotal Loss: 0.0474\tMAELoss: 0.0734\tMSSSIM Loss: 0.0425\n",
      "[0/200][32/89]\tTotal Loss: 0.0361\tMAELoss: 0.0542\tMSSSIM Loss: 0.0326\n",
      "[0/200][33/89]\tTotal Loss: 0.0421\tMAELoss: 0.0385\tMSSSIM Loss: 0.0427\n",
      "[0/200][34/89]\tTotal Loss: 0.0349\tMAELoss: 0.0356\tMSSSIM Loss: 0.0348\n",
      "[0/200][35/89]\tTotal Loss: 0.0370\tMAELoss: 0.0371\tMSSSIM Loss: 0.0370\n",
      "[0/200][36/89]\tTotal Loss: 0.0434\tMAELoss: 0.0429\tMSSSIM Loss: 0.0435\n",
      "[0/200][37/89]\tTotal Loss: 0.0434\tMAELoss: 0.0443\tMSSSIM Loss: 0.0432\n",
      "[0/200][38/89]\tTotal Loss: 0.0343\tMAELoss: 0.0445\tMSSSIM Loss: 0.0323\n",
      "[0/200][39/89]\tTotal Loss: 0.0419\tMAELoss: 0.0424\tMSSSIM Loss: 0.0418\n",
      "[0/200][40/89]\tTotal Loss: 0.0360\tMAELoss: 0.0462\tMSSSIM Loss: 0.0340\n",
      "[0/200][41/89]\tTotal Loss: 0.0435\tMAELoss: 0.0600\tMSSSIM Loss: 0.0404\n",
      "[0/200][42/89]\tTotal Loss: 0.0448\tMAELoss: 0.0625\tMSSSIM Loss: 0.0414\n",
      "[0/200][43/89]\tTotal Loss: 0.0379\tMAELoss: 0.0596\tMSSSIM Loss: 0.0337\n",
      "[0/200][44/89]\tTotal Loss: 0.0506\tMAELoss: 0.0415\tMSSSIM Loss: 0.0523\n",
      "[0/200][45/89]\tTotal Loss: 0.0310\tMAELoss: 0.0328\tMSSSIM Loss: 0.0307\n",
      "[0/200][46/89]\tTotal Loss: 0.0292\tMAELoss: 0.0355\tMSSSIM Loss: 0.0280\n",
      "[0/200][47/89]\tTotal Loss: 0.0393\tMAELoss: 0.0369\tMSSSIM Loss: 0.0397\n",
      "[0/200][48/89]\tTotal Loss: 0.0208\tMAELoss: 0.0310\tMSSSIM Loss: 0.0188\n",
      "[0/200][49/89]\tTotal Loss: 0.0361\tMAELoss: 0.0372\tMSSSIM Loss: 0.0359\n",
      "[0/200][50/89]\tTotal Loss: 0.0294\tMAELoss: 0.0298\tMSSSIM Loss: 0.0293\n",
      "[0/200][51/89]\tTotal Loss: 0.0303\tMAELoss: 0.0290\tMSSSIM Loss: 0.0306\n",
      "[0/200][52/89]\tTotal Loss: 0.0298\tMAELoss: 0.0281\tMSSSIM Loss: 0.0302\n",
      "[0/200][53/89]\tTotal Loss: 0.0297\tMAELoss: 0.0285\tMSSSIM Loss: 0.0300\n",
      "[0/200][54/89]\tTotal Loss: 0.0284\tMAELoss: 0.0290\tMSSSIM Loss: 0.0283\n",
      "[0/200][55/89]\tTotal Loss: 0.0426\tMAELoss: 0.0394\tMSSSIM Loss: 0.0432\n",
      "[0/200][56/89]\tTotal Loss: 0.0289\tMAELoss: 0.0275\tMSSSIM Loss: 0.0291\n",
      "[0/200][57/89]\tTotal Loss: 0.0235\tMAELoss: 0.0264\tMSSSIM Loss: 0.0229\n",
      "[0/200][58/89]\tTotal Loss: 0.0358\tMAELoss: 0.0386\tMSSSIM Loss: 0.0353\n",
      "[0/200][59/89]\tTotal Loss: 0.0368\tMAELoss: 0.0319\tMSSSIM Loss: 0.0377\n",
      "[0/200][60/89]\tTotal Loss: 0.0289\tMAELoss: 0.0314\tMSSSIM Loss: 0.0284\n",
      "[0/200][61/89]\tTotal Loss: 0.0281\tMAELoss: 0.0319\tMSSSIM Loss: 0.0273\n",
      "[0/200][62/89]\tTotal Loss: 0.0369\tMAELoss: 0.0322\tMSSSIM Loss: 0.0378\n",
      "[0/200][63/89]\tTotal Loss: 0.0309\tMAELoss: 0.0286\tMSSSIM Loss: 0.0313\n",
      "[0/200][64/89]\tTotal Loss: 0.0326\tMAELoss: 0.0363\tMSSSIM Loss: 0.0319\n",
      "[0/200][65/89]\tTotal Loss: 0.0344\tMAELoss: 0.0383\tMSSSIM Loss: 0.0336\n",
      "[0/200][66/89]\tTotal Loss: 0.0311\tMAELoss: 0.0389\tMSSSIM Loss: 0.0297\n",
      "[0/200][67/89]\tTotal Loss: 0.0320\tMAELoss: 0.0294\tMSSSIM Loss: 0.0325\n",
      "[0/200][68/89]\tTotal Loss: 0.0271\tMAELoss: 0.0294\tMSSSIM Loss: 0.0266\n",
      "[0/200][69/89]\tTotal Loss: 0.0285\tMAELoss: 0.0234\tMSSSIM Loss: 0.0295\n",
      "[0/200][70/89]\tTotal Loss: 0.0173\tMAELoss: 0.0150\tMSSSIM Loss: 0.0177\n",
      "[0/200][71/89]\tTotal Loss: 0.0386\tMAELoss: 0.0373\tMSSSIM Loss: 0.0388\n",
      "[0/200][72/89]\tTotal Loss: 0.0194\tMAELoss: 0.0278\tMSSSIM Loss: 0.0178\n",
      "[0/200][73/89]\tTotal Loss: 0.0294\tMAELoss: 0.0284\tMSSSIM Loss: 0.0296\n",
      "[0/200][74/89]\tTotal Loss: 0.0257\tMAELoss: 0.0249\tMSSSIM Loss: 0.0258\n",
      "[0/200][75/89]\tTotal Loss: 0.0267\tMAELoss: 0.0221\tMSSSIM Loss: 0.0275\n",
      "[0/200][76/89]\tTotal Loss: 0.0228\tMAELoss: 0.0208\tMSSSIM Loss: 0.0232\n",
      "[0/200][77/89]\tTotal Loss: 0.0252\tMAELoss: 0.0335\tMSSSIM Loss: 0.0237\n",
      "[0/200][78/89]\tTotal Loss: 0.0332\tMAELoss: 0.0383\tMSSSIM Loss: 0.0323\n",
      "[0/200][79/89]\tTotal Loss: 0.0264\tMAELoss: 0.0309\tMSSSIM Loss: 0.0255\n",
      "[0/200][80/89]\tTotal Loss: 0.0226\tMAELoss: 0.0213\tMSSSIM Loss: 0.0229\n",
      "[0/200][81/89]\tTotal Loss: 0.0344\tMAELoss: 0.0291\tMSSSIM Loss: 0.0354\n",
      "[0/200][82/89]\tTotal Loss: 0.0344\tMAELoss: 0.0310\tMSSSIM Loss: 0.0351\n",
      "[0/200][83/89]\tTotal Loss: 0.0358\tMAELoss: 0.0331\tMSSSIM Loss: 0.0363\n",
      "[0/200][84/89]\tTotal Loss: 0.0406\tMAELoss: 0.0323\tMSSSIM Loss: 0.0422\n",
      "[0/200][85/89]\tTotal Loss: 0.0311\tMAELoss: 0.0366\tMSSSIM Loss: 0.0301\n",
      "[0/200][86/89]\tTotal Loss: 0.0171\tMAELoss: 0.0215\tMSSSIM Loss: 0.0163\n",
      "[0/200][87/89]\tTotal Loss: 0.0258\tMAELoss: 0.0278\tMSSSIM Loss: 0.0254\n",
      "[0/200][88/89]\tTotal Loss: 0.0266\tMAELoss: 0.0238\tMSSSIM Loss: 0.0271\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    44: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    56: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    78: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    89: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch   100: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch   116: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   127: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_2.tar\n",
      "Fold 3\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.6980\tMAELoss: 0.2837\tMSSSIM Loss: 0.7769\n",
      "[0/200][1/89]\tTotal Loss: 0.6449\tMAELoss: 0.2768\tMSSSIM Loss: 0.7150\n",
      "[0/200][2/89]\tTotal Loss: 0.5332\tMAELoss: 0.2819\tMSSSIM Loss: 0.5810\n",
      "[0/200][3/89]\tTotal Loss: 0.3222\tMAELoss: 0.2324\tMSSSIM Loss: 0.3394\n",
      "[0/200][4/89]\tTotal Loss: 0.1951\tMAELoss: 0.3798\tMSSSIM Loss: 0.1599\n",
      "[0/200][5/89]\tTotal Loss: 0.4002\tMAELoss: 0.4119\tMSSSIM Loss: 0.3979\n",
      "[0/200][6/89]\tTotal Loss: 0.3701\tMAELoss: 0.4135\tMSSSIM Loss: 0.3619\n",
      "[0/200][7/89]\tTotal Loss: 0.4040\tMAELoss: 0.2659\tMSSSIM Loss: 0.4303\n",
      "[0/200][8/89]\tTotal Loss: 0.1772\tMAELoss: 0.2669\tMSSSIM Loss: 0.1601\n",
      "[0/200][9/89]\tTotal Loss: 0.2080\tMAELoss: 0.2020\tMSSSIM Loss: 0.2091\n",
      "[0/200][10/89]\tTotal Loss: 0.2348\tMAELoss: 0.1619\tMSSSIM Loss: 0.2487\n",
      "[0/200][11/89]\tTotal Loss: 0.1840\tMAELoss: 0.1444\tMSSSIM Loss: 0.1916\n",
      "[0/200][12/89]\tTotal Loss: 0.1636\tMAELoss: 0.0992\tMSSSIM Loss: 0.1759\n",
      "[0/200][13/89]\tTotal Loss: 0.1374\tMAELoss: 0.0966\tMSSSIM Loss: 0.1452\n",
      "[0/200][14/89]\tTotal Loss: 0.1242\tMAELoss: 0.0924\tMSSSIM Loss: 0.1303\n",
      "[0/200][15/89]\tTotal Loss: 0.0852\tMAELoss: 0.0784\tMSSSIM Loss: 0.0865\n",
      "[0/200][16/89]\tTotal Loss: 0.0890\tMAELoss: 0.0864\tMSSSIM Loss: 0.0895\n",
      "[0/200][17/89]\tTotal Loss: 0.0756\tMAELoss: 0.1164\tMSSSIM Loss: 0.0678\n",
      "[0/200][18/89]\tTotal Loss: 0.0609\tMAELoss: 0.0786\tMSSSIM Loss: 0.0576\n",
      "[0/200][19/89]\tTotal Loss: 0.0476\tMAELoss: 0.0428\tMSSSIM Loss: 0.0485\n",
      "[0/200][20/89]\tTotal Loss: 0.0592\tMAELoss: 0.0539\tMSSSIM Loss: 0.0602\n",
      "[0/200][21/89]\tTotal Loss: 0.0604\tMAELoss: 0.0593\tMSSSIM Loss: 0.0606\n",
      "[0/200][22/89]\tTotal Loss: 0.0558\tMAELoss: 0.0468\tMSSSIM Loss: 0.0576\n",
      "[0/200][23/89]\tTotal Loss: 0.0424\tMAELoss: 0.0495\tMSSSIM Loss: 0.0410\n",
      "[0/200][24/89]\tTotal Loss: 0.0467\tMAELoss: 0.1071\tMSSSIM Loss: 0.0351\n",
      "[0/200][25/89]\tTotal Loss: 0.0619\tMAELoss: 0.1088\tMSSSIM Loss: 0.0530\n",
      "[0/200][26/89]\tTotal Loss: 0.0467\tMAELoss: 0.0691\tMSSSIM Loss: 0.0424\n",
      "[0/200][27/89]\tTotal Loss: 0.0431\tMAELoss: 0.0403\tMSSSIM Loss: 0.0436\n",
      "[0/200][28/89]\tTotal Loss: 0.0334\tMAELoss: 0.0267\tMSSSIM Loss: 0.0347\n",
      "[0/200][29/89]\tTotal Loss: 0.0377\tMAELoss: 0.0322\tMSSSIM Loss: 0.0388\n",
      "[0/200][30/89]\tTotal Loss: 0.0594\tMAELoss: 0.0723\tMSSSIM Loss: 0.0569\n",
      "[0/200][31/89]\tTotal Loss: 0.0406\tMAELoss: 0.0656\tMSSSIM Loss: 0.0359\n",
      "[0/200][32/89]\tTotal Loss: 0.0484\tMAELoss: 0.0561\tMSSSIM Loss: 0.0469\n",
      "[0/200][33/89]\tTotal Loss: 0.0452\tMAELoss: 0.0425\tMSSSIM Loss: 0.0457\n",
      "[0/200][34/89]\tTotal Loss: 0.0501\tMAELoss: 0.0617\tMSSSIM Loss: 0.0479\n",
      "[0/200][35/89]\tTotal Loss: 0.0617\tMAELoss: 0.1070\tMSSSIM Loss: 0.0531\n",
      "[0/200][36/89]\tTotal Loss: 0.0529\tMAELoss: 0.1239\tMSSSIM Loss: 0.0394\n",
      "[0/200][37/89]\tTotal Loss: 0.0567\tMAELoss: 0.0732\tMSSSIM Loss: 0.0535\n",
      "[0/200][38/89]\tTotal Loss: 0.0533\tMAELoss: 0.0694\tMSSSIM Loss: 0.0502\n",
      "[0/200][39/89]\tTotal Loss: 0.0256\tMAELoss: 0.0263\tMSSSIM Loss: 0.0255\n",
      "[0/200][40/89]\tTotal Loss: 0.0546\tMAELoss: 0.0664\tMSSSIM Loss: 0.0524\n",
      "[0/200][41/89]\tTotal Loss: 0.0480\tMAELoss: 0.0534\tMSSSIM Loss: 0.0469\n",
      "[0/200][42/89]\tTotal Loss: 0.0556\tMAELoss: 0.0576\tMSSSIM Loss: 0.0552\n",
      "[0/200][43/89]\tTotal Loss: 0.0393\tMAELoss: 0.0520\tMSSSIM Loss: 0.0369\n",
      "[0/200][44/89]\tTotal Loss: 0.0512\tMAELoss: 0.0628\tMSSSIM Loss: 0.0490\n",
      "[0/200][45/89]\tTotal Loss: 0.0443\tMAELoss: 0.0512\tMSSSIM Loss: 0.0430\n",
      "[0/200][46/89]\tTotal Loss: 0.0377\tMAELoss: 0.0381\tMSSSIM Loss: 0.0376\n",
      "[0/200][47/89]\tTotal Loss: 0.0378\tMAELoss: 0.0568\tMSSSIM Loss: 0.0342\n",
      "[0/200][48/89]\tTotal Loss: 0.0352\tMAELoss: 0.0530\tMSSSIM Loss: 0.0318\n",
      "[0/200][49/89]\tTotal Loss: 0.0372\tMAELoss: 0.0375\tMSSSIM Loss: 0.0371\n",
      "[0/200][50/89]\tTotal Loss: 0.0612\tMAELoss: 0.0419\tMSSSIM Loss: 0.0649\n",
      "[0/200][51/89]\tTotal Loss: 0.0296\tMAELoss: 0.0315\tMSSSIM Loss: 0.0293\n",
      "[0/200][52/89]\tTotal Loss: 0.0395\tMAELoss: 0.0547\tMSSSIM Loss: 0.0366\n",
      "[0/200][53/89]\tTotal Loss: 0.0425\tMAELoss: 0.0727\tMSSSIM Loss: 0.0368\n",
      "[0/200][54/89]\tTotal Loss: 0.0418\tMAELoss: 0.0683\tMSSSIM Loss: 0.0368\n",
      "[0/200][55/89]\tTotal Loss: 0.0357\tMAELoss: 0.0447\tMSSSIM Loss: 0.0340\n",
      "[0/200][56/89]\tTotal Loss: 0.0265\tMAELoss: 0.0302\tMSSSIM Loss: 0.0258\n",
      "[0/200][57/89]\tTotal Loss: 0.0286\tMAELoss: 0.0272\tMSSSIM Loss: 0.0288\n",
      "[0/200][58/89]\tTotal Loss: 0.0463\tMAELoss: 0.0403\tMSSSIM Loss: 0.0474\n",
      "[0/200][59/89]\tTotal Loss: 0.0337\tMAELoss: 0.0319\tMSSSIM Loss: 0.0341\n",
      "[0/200][60/89]\tTotal Loss: 0.0340\tMAELoss: 0.0332\tMSSSIM Loss: 0.0342\n",
      "[0/200][61/89]\tTotal Loss: 0.0380\tMAELoss: 0.0432\tMSSSIM Loss: 0.0370\n",
      "[0/200][62/89]\tTotal Loss: 0.0371\tMAELoss: 0.0436\tMSSSIM Loss: 0.0359\n",
      "[0/200][63/89]\tTotal Loss: 0.0312\tMAELoss: 0.0414\tMSSSIM Loss: 0.0292\n",
      "[0/200][64/89]\tTotal Loss: 0.0387\tMAELoss: 0.0486\tMSSSIM Loss: 0.0368\n",
      "[0/200][65/89]\tTotal Loss: 0.0279\tMAELoss: 0.0330\tMSSSIM Loss: 0.0270\n",
      "[0/200][66/89]\tTotal Loss: 0.0451\tMAELoss: 0.0321\tMSSSIM Loss: 0.0476\n",
      "[0/200][67/89]\tTotal Loss: 0.0278\tMAELoss: 0.0278\tMSSSIM Loss: 0.0278\n",
      "[0/200][68/89]\tTotal Loss: 0.0325\tMAELoss: 0.0311\tMSSSIM Loss: 0.0328\n",
      "[0/200][69/89]\tTotal Loss: 0.0341\tMAELoss: 0.0327\tMSSSIM Loss: 0.0344\n",
      "[0/200][70/89]\tTotal Loss: 0.0353\tMAELoss: 0.0401\tMSSSIM Loss: 0.0343\n",
      "[0/200][71/89]\tTotal Loss: 0.0244\tMAELoss: 0.0345\tMSSSIM Loss: 0.0224\n",
      "[0/200][72/89]\tTotal Loss: 0.0405\tMAELoss: 0.0407\tMSSSIM Loss: 0.0405\n",
      "[0/200][73/89]\tTotal Loss: 0.0287\tMAELoss: 0.0313\tMSSSIM Loss: 0.0282\n",
      "[0/200][74/89]\tTotal Loss: 0.0452\tMAELoss: 0.0339\tMSSSIM Loss: 0.0474\n",
      "[0/200][75/89]\tTotal Loss: 0.0357\tMAELoss: 0.0356\tMSSSIM Loss: 0.0357\n",
      "[0/200][76/89]\tTotal Loss: 0.0290\tMAELoss: 0.0347\tMSSSIM Loss: 0.0279\n",
      "[0/200][77/89]\tTotal Loss: 0.0371\tMAELoss: 0.0343\tMSSSIM Loss: 0.0376\n",
      "[0/200][78/89]\tTotal Loss: 0.0294\tMAELoss: 0.0325\tMSSSIM Loss: 0.0288\n",
      "[0/200][79/89]\tTotal Loss: 0.0264\tMAELoss: 0.0279\tMSSSIM Loss: 0.0261\n",
      "[0/200][80/89]\tTotal Loss: 0.0193\tMAELoss: 0.0231\tMSSSIM Loss: 0.0186\n",
      "[0/200][81/89]\tTotal Loss: 0.0310\tMAELoss: 0.0308\tMSSSIM Loss: 0.0310\n",
      "[0/200][82/89]\tTotal Loss: 0.0333\tMAELoss: 0.0297\tMSSSIM Loss: 0.0340\n",
      "[0/200][83/89]\tTotal Loss: 0.0157\tMAELoss: 0.0180\tMSSSIM Loss: 0.0153\n",
      "[0/200][84/89]\tTotal Loss: 0.0245\tMAELoss: 0.0264\tMSSSIM Loss: 0.0241\n",
      "[0/200][85/89]\tTotal Loss: 0.0231\tMAELoss: 0.0234\tMSSSIM Loss: 0.0231\n",
      "[0/200][86/89]\tTotal Loss: 0.0253\tMAELoss: 0.0231\tMSSSIM Loss: 0.0257\n",
      "[0/200][87/89]\tTotal Loss: 0.0241\tMAELoss: 0.0242\tMSSSIM Loss: 0.0241\n",
      "[0/200][88/89]\tTotal Loss: 0.0267\tMAELoss: 0.0246\tMSSSIM Loss: 0.0271\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    25: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    40: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    51: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    68: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    89: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch   100: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   111: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_3.tar\n",
      "Fold 4\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.6922\tMAELoss: 0.3102\tMSSSIM Loss: 0.7650\n",
      "[0/200][1/89]\tTotal Loss: 0.5813\tMAELoss: 0.3098\tMSSSIM Loss: 0.6330\n",
      "[0/200][2/89]\tTotal Loss: 0.5638\tMAELoss: 0.2549\tMSSSIM Loss: 0.6227\n",
      "[0/200][3/89]\tTotal Loss: 0.4402\tMAELoss: 0.2516\tMSSSIM Loss: 0.4761\n",
      "[0/200][4/89]\tTotal Loss: 0.2613\tMAELoss: 0.2715\tMSSSIM Loss: 0.2594\n",
      "[0/200][5/89]\tTotal Loss: 0.3467\tMAELoss: 0.3604\tMSSSIM Loss: 0.3441\n",
      "[0/200][6/89]\tTotal Loss: 0.2896\tMAELoss: 0.4524\tMSSSIM Loss: 0.2586\n",
      "[0/200][7/89]\tTotal Loss: 0.3928\tMAELoss: 0.2374\tMSSSIM Loss: 0.4224\n",
      "[0/200][8/89]\tTotal Loss: 0.2158\tMAELoss: 0.1939\tMSSSIM Loss: 0.2199\n",
      "[0/200][9/89]\tTotal Loss: 0.2517\tMAELoss: 0.1858\tMSSSIM Loss: 0.2643\n",
      "[0/200][10/89]\tTotal Loss: 0.2806\tMAELoss: 0.2022\tMSSSIM Loss: 0.2955\n",
      "[0/200][11/89]\tTotal Loss: 0.1952\tMAELoss: 0.1696\tMSSSIM Loss: 0.2001\n",
      "[0/200][12/89]\tTotal Loss: 0.1317\tMAELoss: 0.1247\tMSSSIM Loss: 0.1330\n",
      "[0/200][13/89]\tTotal Loss: 0.1185\tMAELoss: 0.0893\tMSSSIM Loss: 0.1240\n",
      "[0/200][14/89]\tTotal Loss: 0.1456\tMAELoss: 0.0893\tMSSSIM Loss: 0.1563\n",
      "[0/200][15/89]\tTotal Loss: 0.1163\tMAELoss: 0.0868\tMSSSIM Loss: 0.1219\n",
      "[0/200][16/89]\tTotal Loss: 0.1033\tMAELoss: 0.0789\tMSSSIM Loss: 0.1080\n",
      "[0/200][17/89]\tTotal Loss: 0.0568\tMAELoss: 0.1044\tMSSSIM Loss: 0.0478\n",
      "[0/200][18/89]\tTotal Loss: 0.0683\tMAELoss: 0.0969\tMSSSIM Loss: 0.0628\n",
      "[0/200][19/89]\tTotal Loss: 0.0675\tMAELoss: 0.0819\tMSSSIM Loss: 0.0647\n",
      "[0/200][20/89]\tTotal Loss: 0.0693\tMAELoss: 0.0685\tMSSSIM Loss: 0.0694\n",
      "[0/200][21/89]\tTotal Loss: 0.0541\tMAELoss: 0.0798\tMSSSIM Loss: 0.0492\n",
      "[0/200][22/89]\tTotal Loss: 0.0560\tMAELoss: 0.0594\tMSSSIM Loss: 0.0554\n",
      "[0/200][23/89]\tTotal Loss: 0.0559\tMAELoss: 0.0556\tMSSSIM Loss: 0.0559\n",
      "[0/200][24/89]\tTotal Loss: 0.0589\tMAELoss: 0.0638\tMSSSIM Loss: 0.0580\n",
      "[0/200][25/89]\tTotal Loss: 0.0400\tMAELoss: 0.0466\tMSSSIM Loss: 0.0387\n",
      "[0/200][26/89]\tTotal Loss: 0.0526\tMAELoss: 0.0445\tMSSSIM Loss: 0.0541\n",
      "[0/200][27/89]\tTotal Loss: 0.0545\tMAELoss: 0.0468\tMSSSIM Loss: 0.0559\n",
      "[0/200][28/89]\tTotal Loss: 0.0397\tMAELoss: 0.0402\tMSSSIM Loss: 0.0396\n",
      "[0/200][29/89]\tTotal Loss: 0.0389\tMAELoss: 0.0389\tMSSSIM Loss: 0.0389\n",
      "[0/200][30/89]\tTotal Loss: 0.0397\tMAELoss: 0.0548\tMSSSIM Loss: 0.0368\n",
      "[0/200][31/89]\tTotal Loss: 0.0453\tMAELoss: 0.0691\tMSSSIM Loss: 0.0408\n",
      "[0/200][32/89]\tTotal Loss: 0.0395\tMAELoss: 0.0586\tMSSSIM Loss: 0.0358\n",
      "[0/200][33/89]\tTotal Loss: 0.0350\tMAELoss: 0.0469\tMSSSIM Loss: 0.0327\n",
      "[0/200][34/89]\tTotal Loss: 0.0410\tMAELoss: 0.0476\tMSSSIM Loss: 0.0397\n",
      "[0/200][35/89]\tTotal Loss: 0.0345\tMAELoss: 0.0410\tMSSSIM Loss: 0.0332\n",
      "[0/200][36/89]\tTotal Loss: 0.0450\tMAELoss: 0.0496\tMSSSIM Loss: 0.0442\n",
      "[0/200][37/89]\tTotal Loss: 0.0359\tMAELoss: 0.0304\tMSSSIM Loss: 0.0370\n",
      "[0/200][38/89]\tTotal Loss: 0.0420\tMAELoss: 0.0396\tMSSSIM Loss: 0.0424\n",
      "[0/200][39/89]\tTotal Loss: 0.0290\tMAELoss: 0.0516\tMSSSIM Loss: 0.0247\n",
      "[0/200][40/89]\tTotal Loss: 0.0298\tMAELoss: 0.0425\tMSSSIM Loss: 0.0274\n",
      "[0/200][41/89]\tTotal Loss: 0.0320\tMAELoss: 0.0375\tMSSSIM Loss: 0.0310\n",
      "[0/200][42/89]\tTotal Loss: 0.0519\tMAELoss: 0.0382\tMSSSIM Loss: 0.0545\n",
      "[0/200][43/89]\tTotal Loss: 0.0362\tMAELoss: 0.0537\tMSSSIM Loss: 0.0328\n",
      "[0/200][44/89]\tTotal Loss: 0.0400\tMAELoss: 0.0770\tMSSSIM Loss: 0.0330\n",
      "[0/200][45/89]\tTotal Loss: 0.0348\tMAELoss: 0.0788\tMSSSIM Loss: 0.0264\n",
      "[0/200][46/89]\tTotal Loss: 0.0362\tMAELoss: 0.0364\tMSSSIM Loss: 0.0362\n",
      "[0/200][47/89]\tTotal Loss: 0.0345\tMAELoss: 0.0315\tMSSSIM Loss: 0.0351\n",
      "[0/200][48/89]\tTotal Loss: 0.0487\tMAELoss: 0.0386\tMSSSIM Loss: 0.0506\n",
      "[0/200][49/89]\tTotal Loss: 0.0308\tMAELoss: 0.0280\tMSSSIM Loss: 0.0313\n",
      "[0/200][50/89]\tTotal Loss: 0.0276\tMAELoss: 0.0315\tMSSSIM Loss: 0.0268\n",
      "[0/200][51/89]\tTotal Loss: 0.0412\tMAELoss: 0.0477\tMSSSIM Loss: 0.0399\n",
      "[0/200][52/89]\tTotal Loss: 0.0365\tMAELoss: 0.0501\tMSSSIM Loss: 0.0340\n",
      "[0/200][53/89]\tTotal Loss: 0.0249\tMAELoss: 0.0270\tMSSSIM Loss: 0.0245\n",
      "[0/200][54/89]\tTotal Loss: 0.0354\tMAELoss: 0.0314\tMSSSIM Loss: 0.0362\n",
      "[0/200][55/89]\tTotal Loss: 0.0312\tMAELoss: 0.0324\tMSSSIM Loss: 0.0310\n",
      "[0/200][56/89]\tTotal Loss: 0.0389\tMAELoss: 0.0340\tMSSSIM Loss: 0.0398\n",
      "[0/200][57/89]\tTotal Loss: 0.0273\tMAELoss: 0.0381\tMSSSIM Loss: 0.0252\n",
      "[0/200][58/89]\tTotal Loss: 0.0270\tMAELoss: 0.0289\tMSSSIM Loss: 0.0266\n",
      "[0/200][59/89]\tTotal Loss: 0.0295\tMAELoss: 0.0295\tMSSSIM Loss: 0.0296\n",
      "[0/200][60/89]\tTotal Loss: 0.0181\tMAELoss: 0.0183\tMSSSIM Loss: 0.0181\n",
      "[0/200][61/89]\tTotal Loss: 0.0288\tMAELoss: 0.0258\tMSSSIM Loss: 0.0294\n",
      "[0/200][62/89]\tTotal Loss: 0.0264\tMAELoss: 0.0330\tMSSSIM Loss: 0.0251\n",
      "[0/200][63/89]\tTotal Loss: 0.0299\tMAELoss: 0.0434\tMSSSIM Loss: 0.0274\n",
      "[0/200][64/89]\tTotal Loss: 0.0273\tMAELoss: 0.0309\tMSSSIM Loss: 0.0267\n",
      "[0/200][65/89]\tTotal Loss: 0.0316\tMAELoss: 0.0321\tMSSSIM Loss: 0.0315\n",
      "[0/200][66/89]\tTotal Loss: 0.0263\tMAELoss: 0.0246\tMSSSIM Loss: 0.0266\n",
      "[0/200][67/89]\tTotal Loss: 0.0288\tMAELoss: 0.0255\tMSSSIM Loss: 0.0295\n",
      "[0/200][68/89]\tTotal Loss: 0.0337\tMAELoss: 0.0320\tMSSSIM Loss: 0.0340\n",
      "[0/200][69/89]\tTotal Loss: 0.0319\tMAELoss: 0.0339\tMSSSIM Loss: 0.0315\n",
      "[0/200][70/89]\tTotal Loss: 0.0232\tMAELoss: 0.0283\tMSSSIM Loss: 0.0222\n",
      "[0/200][71/89]\tTotal Loss: 0.0280\tMAELoss: 0.0250\tMSSSIM Loss: 0.0286\n",
      "[0/200][72/89]\tTotal Loss: 0.0249\tMAELoss: 0.0271\tMSSSIM Loss: 0.0245\n",
      "[0/200][73/89]\tTotal Loss: 0.0445\tMAELoss: 0.0513\tMSSSIM Loss: 0.0432\n",
      "[0/200][74/89]\tTotal Loss: 0.0219\tMAELoss: 0.0244\tMSSSIM Loss: 0.0214\n",
      "[0/200][75/89]\tTotal Loss: 0.0336\tMAELoss: 0.0439\tMSSSIM Loss: 0.0316\n",
      "[0/200][76/89]\tTotal Loss: 0.0371\tMAELoss: 0.0401\tMSSSIM Loss: 0.0365\n",
      "[0/200][77/89]\tTotal Loss: 0.0530\tMAELoss: 0.0770\tMSSSIM Loss: 0.0485\n",
      "[0/200][78/89]\tTotal Loss: 0.0341\tMAELoss: 0.0571\tMSSSIM Loss: 0.0297\n",
      "[0/200][79/89]\tTotal Loss: 0.0484\tMAELoss: 0.0529\tMSSSIM Loss: 0.0475\n",
      "[0/200][80/89]\tTotal Loss: 0.0379\tMAELoss: 0.0344\tMSSSIM Loss: 0.0386\n",
      "[0/200][81/89]\tTotal Loss: 0.0282\tMAELoss: 0.0319\tMSSSIM Loss: 0.0275\n",
      "[0/200][82/89]\tTotal Loss: 0.0306\tMAELoss: 0.0311\tMSSSIM Loss: 0.0305\n",
      "[0/200][83/89]\tTotal Loss: 0.0425\tMAELoss: 0.0327\tMSSSIM Loss: 0.0444\n",
      "[0/200][84/89]\tTotal Loss: 0.0331\tMAELoss: 0.0333\tMSSSIM Loss: 0.0331\n",
      "[0/200][85/89]\tTotal Loss: 0.0358\tMAELoss: 0.0452\tMSSSIM Loss: 0.0340\n",
      "[0/200][86/89]\tTotal Loss: 0.0466\tMAELoss: 0.0549\tMSSSIM Loss: 0.0450\n",
      "[0/200][87/89]\tTotal Loss: 0.0377\tMAELoss: 0.0481\tMSSSIM Loss: 0.0358\n",
      "[0/200][88/89]\tTotal Loss: 0.0327\tMAELoss: 0.0339\tMSSSIM Loss: 0.0325\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    20: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    31: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    42: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    56: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    67: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch    78: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch    96: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_4.tar\n",
      "Fold 5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.8861\tMAELoss: 0.2882\tMSSSIM Loss: 1.0000\n",
      "[0/200][1/89]\tTotal Loss: 0.6624\tMAELoss: 0.2821\tMSSSIM Loss: 0.7348\n",
      "[0/200][2/89]\tTotal Loss: 0.5654\tMAELoss: 0.2738\tMSSSIM Loss: 0.6209\n",
      "[0/200][3/89]\tTotal Loss: 0.3911\tMAELoss: 0.1762\tMSSSIM Loss: 0.4320\n",
      "[0/200][4/89]\tTotal Loss: 0.2113\tMAELoss: 0.3910\tMSSSIM Loss: 0.1770\n",
      "[0/200][5/89]\tTotal Loss: 0.3726\tMAELoss: 0.4072\tMSSSIM Loss: 0.3660\n",
      "[0/200][6/89]\tTotal Loss: 0.3980\tMAELoss: 0.4169\tMSSSIM Loss: 0.3944\n",
      "[0/200][7/89]\tTotal Loss: 0.3470\tMAELoss: 0.3348\tMSSSIM Loss: 0.3493\n",
      "[0/200][8/89]\tTotal Loss: 0.1937\tMAELoss: 0.3141\tMSSSIM Loss: 0.1708\n",
      "[0/200][9/89]\tTotal Loss: 0.2148\tMAELoss: 0.2509\tMSSSIM Loss: 0.2080\n",
      "[0/200][10/89]\tTotal Loss: 0.1759\tMAELoss: 0.2358\tMSSSIM Loss: 0.1645\n",
      "[0/200][11/89]\tTotal Loss: 0.1450\tMAELoss: 0.1561\tMSSSIM Loss: 0.1429\n",
      "[0/200][12/89]\tTotal Loss: 0.1178\tMAELoss: 0.1052\tMSSSIM Loss: 0.1202\n",
      "[0/200][13/89]\tTotal Loss: 0.1120\tMAELoss: 0.0863\tMSSSIM Loss: 0.1169\n",
      "[0/200][14/89]\tTotal Loss: 0.1217\tMAELoss: 0.0950\tMSSSIM Loss: 0.1268\n",
      "[0/200][15/89]\tTotal Loss: 0.3058\tMAELoss: 0.1875\tMSSSIM Loss: 0.3283\n",
      "[0/200][16/89]\tTotal Loss: 0.0637\tMAELoss: 0.1336\tMSSSIM Loss: 0.0504\n",
      "[0/200][17/89]\tTotal Loss: 0.0543\tMAELoss: 0.0685\tMSSSIM Loss: 0.0516\n",
      "[0/200][18/89]\tTotal Loss: 0.0616\tMAELoss: 0.0942\tMSSSIM Loss: 0.0554\n",
      "[0/200][19/89]\tTotal Loss: 0.0684\tMAELoss: 0.1235\tMSSSIM Loss: 0.0579\n",
      "[0/200][20/89]\tTotal Loss: 0.0988\tMAELoss: 0.1318\tMSSSIM Loss: 0.0925\n",
      "[0/200][21/89]\tTotal Loss: 0.0686\tMAELoss: 0.1305\tMSSSIM Loss: 0.0568\n",
      "[0/200][22/89]\tTotal Loss: 0.0879\tMAELoss: 0.1110\tMSSSIM Loss: 0.0835\n",
      "[0/200][23/89]\tTotal Loss: 0.0674\tMAELoss: 0.0931\tMSSSIM Loss: 0.0625\n",
      "[0/200][24/89]\tTotal Loss: 0.0605\tMAELoss: 0.0745\tMSSSIM Loss: 0.0578\n",
      "[0/200][25/89]\tTotal Loss: 0.0496\tMAELoss: 0.0492\tMSSSIM Loss: 0.0496\n",
      "[0/200][26/89]\tTotal Loss: 0.0670\tMAELoss: 0.0669\tMSSSIM Loss: 0.0671\n",
      "[0/200][27/89]\tTotal Loss: 0.0554\tMAELoss: 0.0632\tMSSSIM Loss: 0.0539\n",
      "[0/200][28/89]\tTotal Loss: 0.0406\tMAELoss: 0.0516\tMSSSIM Loss: 0.0385\n",
      "[0/200][29/89]\tTotal Loss: 0.0473\tMAELoss: 0.0448\tMSSSIM Loss: 0.0477\n",
      "[0/200][30/89]\tTotal Loss: 0.0586\tMAELoss: 0.0492\tMSSSIM Loss: 0.0604\n",
      "[0/200][31/89]\tTotal Loss: 0.0557\tMAELoss: 0.0726\tMSSSIM Loss: 0.0524\n",
      "[0/200][32/89]\tTotal Loss: 0.0418\tMAELoss: 0.0461\tMSSSIM Loss: 0.0409\n",
      "[0/200][33/89]\tTotal Loss: 0.0587\tMAELoss: 0.0716\tMSSSIM Loss: 0.0562\n",
      "[0/200][34/89]\tTotal Loss: 0.0535\tMAELoss: 0.0584\tMSSSIM Loss: 0.0525\n",
      "[0/200][35/89]\tTotal Loss: 0.0441\tMAELoss: 0.0451\tMSSSIM Loss: 0.0439\n",
      "[0/200][36/89]\tTotal Loss: 0.0447\tMAELoss: 0.0430\tMSSSIM Loss: 0.0451\n",
      "[0/200][37/89]\tTotal Loss: 0.0313\tMAELoss: 0.0298\tMSSSIM Loss: 0.0316\n",
      "[0/200][38/89]\tTotal Loss: 0.0323\tMAELoss: 0.0401\tMSSSIM Loss: 0.0308\n",
      "[0/200][39/89]\tTotal Loss: 0.0407\tMAELoss: 0.0384\tMSSSIM Loss: 0.0412\n",
      "[0/200][40/89]\tTotal Loss: 0.0384\tMAELoss: 0.0425\tMSSSIM Loss: 0.0376\n",
      "[0/200][41/89]\tTotal Loss: 0.0484\tMAELoss: 0.0518\tMSSSIM Loss: 0.0477\n",
      "[0/200][42/89]\tTotal Loss: 0.0378\tMAELoss: 0.0390\tMSSSIM Loss: 0.0375\n",
      "[0/200][43/89]\tTotal Loss: 0.0618\tMAELoss: 0.0519\tMSSSIM Loss: 0.0637\n",
      "[0/200][44/89]\tTotal Loss: 0.0393\tMAELoss: 0.0451\tMSSSIM Loss: 0.0382\n",
      "[0/200][45/89]\tTotal Loss: 0.0360\tMAELoss: 0.0385\tMSSSIM Loss: 0.0356\n",
      "[0/200][46/89]\tTotal Loss: 0.0341\tMAELoss: 0.0327\tMSSSIM Loss: 0.0344\n",
      "[0/200][47/89]\tTotal Loss: 0.0419\tMAELoss: 0.0446\tMSSSIM Loss: 0.0414\n",
      "[0/200][48/89]\tTotal Loss: 0.0498\tMAELoss: 0.0539\tMSSSIM Loss: 0.0491\n",
      "[0/200][49/89]\tTotal Loss: 0.0283\tMAELoss: 0.0327\tMSSSIM Loss: 0.0275\n",
      "[0/200][50/89]\tTotal Loss: 0.0466\tMAELoss: 0.0410\tMSSSIM Loss: 0.0477\n",
      "[0/200][51/89]\tTotal Loss: 0.0297\tMAELoss: 0.0383\tMSSSIM Loss: 0.0281\n",
      "[0/200][52/89]\tTotal Loss: 0.0342\tMAELoss: 0.0463\tMSSSIM Loss: 0.0319\n",
      "[0/200][53/89]\tTotal Loss: 0.0359\tMAELoss: 0.0384\tMSSSIM Loss: 0.0355\n",
      "[0/200][54/89]\tTotal Loss: 0.0250\tMAELoss: 0.0232\tMSSSIM Loss: 0.0253\n",
      "[0/200][55/89]\tTotal Loss: 0.0246\tMAELoss: 0.0252\tMSSSIM Loss: 0.0245\n",
      "[0/200][56/89]\tTotal Loss: 0.0283\tMAELoss: 0.0324\tMSSSIM Loss: 0.0275\n",
      "[0/200][57/89]\tTotal Loss: 0.0347\tMAELoss: 0.0376\tMSSSIM Loss: 0.0342\n",
      "[0/200][58/89]\tTotal Loss: 0.0276\tMAELoss: 0.0244\tMSSSIM Loss: 0.0282\n",
      "[0/200][59/89]\tTotal Loss: 0.0469\tMAELoss: 0.0455\tMSSSIM Loss: 0.0472\n",
      "[0/200][60/89]\tTotal Loss: 0.0368\tMAELoss: 0.0351\tMSSSIM Loss: 0.0372\n",
      "[0/200][61/89]\tTotal Loss: 0.0246\tMAELoss: 0.0300\tMSSSIM Loss: 0.0235\n",
      "[0/200][62/89]\tTotal Loss: 0.0270\tMAELoss: 0.0326\tMSSSIM Loss: 0.0259\n",
      "[0/200][63/89]\tTotal Loss: 0.0252\tMAELoss: 0.0279\tMSSSIM Loss: 0.0246\n",
      "[0/200][64/89]\tTotal Loss: 0.0325\tMAELoss: 0.0307\tMSSSIM Loss: 0.0329\n",
      "[0/200][65/89]\tTotal Loss: 0.0307\tMAELoss: 0.0296\tMSSSIM Loss: 0.0309\n",
      "[0/200][66/89]\tTotal Loss: 0.0293\tMAELoss: 0.0419\tMSSSIM Loss: 0.0269\n",
      "[0/200][67/89]\tTotal Loss: 0.0281\tMAELoss: 0.0417\tMSSSIM Loss: 0.0255\n",
      "[0/200][68/89]\tTotal Loss: 0.0260\tMAELoss: 0.0296\tMSSSIM Loss: 0.0253\n",
      "[0/200][69/89]\tTotal Loss: 0.0299\tMAELoss: 0.0284\tMSSSIM Loss: 0.0302\n",
      "[0/200][70/89]\tTotal Loss: 0.0345\tMAELoss: 0.0253\tMSSSIM Loss: 0.0362\n",
      "[0/200][71/89]\tTotal Loss: 0.0374\tMAELoss: 0.0437\tMSSSIM Loss: 0.0362\n",
      "[0/200][72/89]\tTotal Loss: 0.0323\tMAELoss: 0.0419\tMSSSIM Loss: 0.0304\n",
      "[0/200][73/89]\tTotal Loss: 0.0165\tMAELoss: 0.0184\tMSSSIM Loss: 0.0162\n",
      "[0/200][74/89]\tTotal Loss: 0.0289\tMAELoss: 0.0267\tMSSSIM Loss: 0.0293\n",
      "[0/200][75/89]\tTotal Loss: 0.0248\tMAELoss: 0.0294\tMSSSIM Loss: 0.0240\n",
      "[0/200][76/89]\tTotal Loss: 0.0374\tMAELoss: 0.0587\tMSSSIM Loss: 0.0333\n",
      "[0/200][77/89]\tTotal Loss: 0.0399\tMAELoss: 0.0505\tMSSSIM Loss: 0.0378\n",
      "[0/200][78/89]\tTotal Loss: 0.0282\tMAELoss: 0.0293\tMSSSIM Loss: 0.0280\n",
      "[0/200][79/89]\tTotal Loss: 0.0304\tMAELoss: 0.0326\tMSSSIM Loss: 0.0300\n",
      "[0/200][80/89]\tTotal Loss: 0.0290\tMAELoss: 0.0295\tMSSSIM Loss: 0.0289\n",
      "[0/200][81/89]\tTotal Loss: 0.0247\tMAELoss: 0.0312\tMSSSIM Loss: 0.0235\n",
      "[0/200][82/89]\tTotal Loss: 0.0241\tMAELoss: 0.0322\tMSSSIM Loss: 0.0225\n",
      "[0/200][83/89]\tTotal Loss: 0.0389\tMAELoss: 0.0402\tMSSSIM Loss: 0.0387\n",
      "[0/200][84/89]\tTotal Loss: 0.0192\tMAELoss: 0.0233\tMSSSIM Loss: 0.0184\n",
      "[0/200][85/89]\tTotal Loss: 0.0331\tMAELoss: 0.0388\tMSSSIM Loss: 0.0320\n",
      "[0/200][86/89]\tTotal Loss: 0.0286\tMAELoss: 0.0303\tMSSSIM Loss: 0.0283\n",
      "[0/200][87/89]\tTotal Loss: 0.0432\tMAELoss: 0.0369\tMSSSIM Loss: 0.0444\n",
      "[0/200][88/89]\tTotal Loss: 0.0265\tMAELoss: 0.0274\tMSSSIM Loss: 0.0263\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    34: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    45: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    65: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    78: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    89: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch   126: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   137: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_5.tar\n",
      "Fold 6\n",
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.7860\tMAELoss: 0.2895\tMSSSIM Loss: 0.8806\n",
      "[0/200][1/89]\tTotal Loss: 0.6621\tMAELoss: 0.2695\tMSSSIM Loss: 0.7369\n",
      "[0/200][2/89]\tTotal Loss: 0.4872\tMAELoss: 0.2260\tMSSSIM Loss: 0.5369\n",
      "[0/200][3/89]\tTotal Loss: 0.3468\tMAELoss: 0.2581\tMSSSIM Loss: 0.3637\n",
      "[0/200][4/89]\tTotal Loss: 0.2430\tMAELoss: 0.2864\tMSSSIM Loss: 0.2347\n",
      "[0/200][5/89]\tTotal Loss: 0.4528\tMAELoss: 0.3018\tMSSSIM Loss: 0.4815\n",
      "[0/200][6/89]\tTotal Loss: 0.3708\tMAELoss: 0.3153\tMSSSIM Loss: 0.3814\n",
      "[0/200][7/89]\tTotal Loss: 0.2439\tMAELoss: 0.3080\tMSSSIM Loss: 0.2317\n",
      "[0/200][8/89]\tTotal Loss: 0.2004\tMAELoss: 0.2697\tMSSSIM Loss: 0.1873\n",
      "[0/200][9/89]\tTotal Loss: 0.2124\tMAELoss: 0.1372\tMSSSIM Loss: 0.2267\n",
      "[0/200][10/89]\tTotal Loss: 0.1658\tMAELoss: 0.1898\tMSSSIM Loss: 0.1613\n",
      "[0/200][11/89]\tTotal Loss: 0.1371\tMAELoss: 0.1393\tMSSSIM Loss: 0.1367\n",
      "[0/200][12/89]\tTotal Loss: 0.1387\tMAELoss: 0.0845\tMSSSIM Loss: 0.1490\n",
      "[0/200][13/89]\tTotal Loss: 0.1143\tMAELoss: 0.0822\tMSSSIM Loss: 0.1205\n",
      "[0/200][14/89]\tTotal Loss: 0.1094\tMAELoss: 0.0956\tMSSSIM Loss: 0.1120\n",
      "[0/200][15/89]\tTotal Loss: 0.2112\tMAELoss: 0.1405\tMSSSIM Loss: 0.2247\n",
      "[0/200][16/89]\tTotal Loss: 0.0865\tMAELoss: 0.0755\tMSSSIM Loss: 0.0885\n",
      "[0/200][17/89]\tTotal Loss: 0.0389\tMAELoss: 0.0347\tMSSSIM Loss: 0.0397\n",
      "[0/200][18/89]\tTotal Loss: 0.0672\tMAELoss: 0.0934\tMSSSIM Loss: 0.0622\n",
      "[0/200][19/89]\tTotal Loss: 0.0741\tMAELoss: 0.1214\tMSSSIM Loss: 0.0651\n",
      "[0/200][20/89]\tTotal Loss: 0.0800\tMAELoss: 0.1133\tMSSSIM Loss: 0.0737\n",
      "[0/200][21/89]\tTotal Loss: 0.0787\tMAELoss: 0.0995\tMSSSIM Loss: 0.0747\n",
      "[0/200][22/89]\tTotal Loss: 0.0789\tMAELoss: 0.1147\tMSSSIM Loss: 0.0720\n",
      "[0/200][23/89]\tTotal Loss: 0.0764\tMAELoss: 0.1009\tMSSSIM Loss: 0.0717\n",
      "[0/200][24/89]\tTotal Loss: 0.0660\tMAELoss: 0.0792\tMSSSIM Loss: 0.0635\n",
      "[0/200][25/89]\tTotal Loss: 0.0712\tMAELoss: 0.0728\tMSSSIM Loss: 0.0709\n",
      "[0/200][26/89]\tTotal Loss: 0.0440\tMAELoss: 0.0508\tMSSSIM Loss: 0.0427\n",
      "[0/200][27/89]\tTotal Loss: 0.0480\tMAELoss: 0.0425\tMSSSIM Loss: 0.0490\n",
      "[0/200][28/89]\tTotal Loss: 0.0409\tMAELoss: 0.0350\tMSSSIM Loss: 0.0420\n",
      "[0/200][29/89]\tTotal Loss: 0.0340\tMAELoss: 0.0782\tMSSSIM Loss: 0.0256\n",
      "[0/200][30/89]\tTotal Loss: 0.0423\tMAELoss: 0.0669\tMSSSIM Loss: 0.0376\n",
      "[0/200][31/89]\tTotal Loss: 0.0401\tMAELoss: 0.0584\tMSSSIM Loss: 0.0366\n",
      "[0/200][32/89]\tTotal Loss: 0.0518\tMAELoss: 0.0465\tMSSSIM Loss: 0.0528\n",
      "[0/200][33/89]\tTotal Loss: 0.0354\tMAELoss: 0.0529\tMSSSIM Loss: 0.0321\n",
      "[0/200][34/89]\tTotal Loss: 0.0382\tMAELoss: 0.0470\tMSSSIM Loss: 0.0365\n",
      "[0/200][35/89]\tTotal Loss: 0.0299\tMAELoss: 0.0497\tMSSSIM Loss: 0.0261\n",
      "[0/200][36/89]\tTotal Loss: 0.0368\tMAELoss: 0.0408\tMSSSIM Loss: 0.0361\n",
      "[0/200][37/89]\tTotal Loss: 0.0432\tMAELoss: 0.0338\tMSSSIM Loss: 0.0449\n",
      "[0/200][38/89]\tTotal Loss: 0.0292\tMAELoss: 0.0366\tMSSSIM Loss: 0.0278\n",
      "[0/200][39/89]\tTotal Loss: 0.0386\tMAELoss: 0.0437\tMSSSIM Loss: 0.0376\n",
      "[0/200][40/89]\tTotal Loss: 0.0414\tMAELoss: 0.0512\tMSSSIM Loss: 0.0396\n",
      "[0/200][41/89]\tTotal Loss: 0.0360\tMAELoss: 0.0376\tMSSSIM Loss: 0.0357\n",
      "[0/200][42/89]\tTotal Loss: 0.0327\tMAELoss: 0.0330\tMSSSIM Loss: 0.0327\n",
      "[0/200][43/89]\tTotal Loss: 0.0342\tMAELoss: 0.0408\tMSSSIM Loss: 0.0329\n",
      "[0/200][44/89]\tTotal Loss: 0.0378\tMAELoss: 0.0474\tMSSSIM Loss: 0.0360\n",
      "[0/200][45/89]\tTotal Loss: 0.0359\tMAELoss: 0.0346\tMSSSIM Loss: 0.0362\n",
      "[0/200][46/89]\tTotal Loss: 0.0301\tMAELoss: 0.0330\tMSSSIM Loss: 0.0295\n",
      "[0/200][47/89]\tTotal Loss: 0.0338\tMAELoss: 0.0298\tMSSSIM Loss: 0.0346\n",
      "[0/200][48/89]\tTotal Loss: 0.0288\tMAELoss: 0.0287\tMSSSIM Loss: 0.0289\n",
      "[0/200][49/89]\tTotal Loss: 0.0428\tMAELoss: 0.0329\tMSSSIM Loss: 0.0447\n",
      "[0/200][50/89]\tTotal Loss: 0.0420\tMAELoss: 0.0599\tMSSSIM Loss: 0.0386\n",
      "[0/200][51/89]\tTotal Loss: 0.0382\tMAELoss: 0.0639\tMSSSIM Loss: 0.0333\n",
      "[0/200][52/89]\tTotal Loss: 0.0250\tMAELoss: 0.0292\tMSSSIM Loss: 0.0242\n",
      "[0/200][53/89]\tTotal Loss: 0.0366\tMAELoss: 0.0392\tMSSSIM Loss: 0.0361\n",
      "[0/200][54/89]\tTotal Loss: 0.0316\tMAELoss: 0.0285\tMSSSIM Loss: 0.0322\n",
      "[0/200][55/89]\tTotal Loss: 0.0449\tMAELoss: 0.0432\tMSSSIM Loss: 0.0453\n",
      "[0/200][56/89]\tTotal Loss: 0.0315\tMAELoss: 0.0295\tMSSSIM Loss: 0.0319\n",
      "[0/200][57/89]\tTotal Loss: 0.0345\tMAELoss: 0.0441\tMSSSIM Loss: 0.0327\n",
      "[0/200][58/89]\tTotal Loss: 0.0348\tMAELoss: 0.0548\tMSSSIM Loss: 0.0310\n",
      "[0/200][59/89]\tTotal Loss: 0.0302\tMAELoss: 0.0301\tMSSSIM Loss: 0.0302\n",
      "[0/200][60/89]\tTotal Loss: 0.0329\tMAELoss: 0.0294\tMSSSIM Loss: 0.0336\n",
      "[0/200][61/89]\tTotal Loss: 0.0527\tMAELoss: 0.0460\tMSSSIM Loss: 0.0539\n",
      "[0/200][62/89]\tTotal Loss: 0.0358\tMAELoss: 0.0412\tMSSSIM Loss: 0.0347\n",
      "[0/200][63/89]\tTotal Loss: 0.0572\tMAELoss: 0.0883\tMSSSIM Loss: 0.0512\n",
      "[0/200][64/89]\tTotal Loss: 0.0486\tMAELoss: 0.0833\tMSSSIM Loss: 0.0419\n",
      "[0/200][65/89]\tTotal Loss: 0.0348\tMAELoss: 0.0460\tMSSSIM Loss: 0.0327\n",
      "[0/200][66/89]\tTotal Loss: 0.0226\tMAELoss: 0.0282\tMSSSIM Loss: 0.0215\n",
      "[0/200][67/89]\tTotal Loss: 0.0382\tMAELoss: 0.0333\tMSSSIM Loss: 0.0391\n",
      "[0/200][68/89]\tTotal Loss: 0.0423\tMAELoss: 0.0406\tMSSSIM Loss: 0.0427\n",
      "[0/200][69/89]\tTotal Loss: 0.0420\tMAELoss: 0.0383\tMSSSIM Loss: 0.0427\n",
      "[0/200][70/89]\tTotal Loss: 0.0300\tMAELoss: 0.0402\tMSSSIM Loss: 0.0281\n",
      "[0/200][71/89]\tTotal Loss: 0.0390\tMAELoss: 0.0472\tMSSSIM Loss: 0.0374\n",
      "[0/200][72/89]\tTotal Loss: 0.0296\tMAELoss: 0.0371\tMSSSIM Loss: 0.0282\n",
      "[0/200][73/89]\tTotal Loss: 0.0292\tMAELoss: 0.0268\tMSSSIM Loss: 0.0296\n",
      "[0/200][74/89]\tTotal Loss: 0.0382\tMAELoss: 0.0300\tMSSSIM Loss: 0.0398\n",
      "[0/200][75/89]\tTotal Loss: 0.0350\tMAELoss: 0.0358\tMSSSIM Loss: 0.0348\n",
      "[0/200][76/89]\tTotal Loss: 0.0212\tMAELoss: 0.0233\tMSSSIM Loss: 0.0208\n",
      "[0/200][77/89]\tTotal Loss: 0.0301\tMAELoss: 0.0297\tMSSSIM Loss: 0.0302\n",
      "[0/200][78/89]\tTotal Loss: 0.0360\tMAELoss: 0.0360\tMSSSIM Loss: 0.0360\n",
      "[0/200][79/89]\tTotal Loss: 0.0321\tMAELoss: 0.0367\tMSSSIM Loss: 0.0312\n",
      "[0/200][80/89]\tTotal Loss: 0.0291\tMAELoss: 0.0320\tMSSSIM Loss: 0.0286\n",
      "[0/200][81/89]\tTotal Loss: 0.0181\tMAELoss: 0.0199\tMSSSIM Loss: 0.0177\n",
      "[0/200][82/89]\tTotal Loss: 0.0329\tMAELoss: 0.0291\tMSSSIM Loss: 0.0336\n",
      "[0/200][83/89]\tTotal Loss: 0.0312\tMAELoss: 0.0311\tMSSSIM Loss: 0.0312\n",
      "[0/200][84/89]\tTotal Loss: 0.0260\tMAELoss: 0.0337\tMSSSIM Loss: 0.0245\n",
      "[0/200][85/89]\tTotal Loss: 0.0241\tMAELoss: 0.0338\tMSSSIM Loss: 0.0222\n",
      "[0/200][86/89]\tTotal Loss: 0.0205\tMAELoss: 0.0293\tMSSSIM Loss: 0.0189\n",
      "[0/200][87/89]\tTotal Loss: 0.0299\tMAELoss: 0.0237\tMSSSIM Loss: 0.0311\n",
      "[0/200][88/89]\tTotal Loss: 0.0426\tMAELoss: 0.0456\tMSSSIM Loss: 0.0420\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    31: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    47: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    58: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    69: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    87: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch    98: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   119: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_6.tar\n",
      "Fold 7\n",
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.6431\tMAELoss: 0.3471\tMSSSIM Loss: 0.6995\n",
      "[0/200][1/89]\tTotal Loss: 0.6086\tMAELoss: 0.2959\tMSSSIM Loss: 0.6682\n",
      "[0/200][2/89]\tTotal Loss: 0.6159\tMAELoss: 0.2759\tMSSSIM Loss: 0.6806\n",
      "[0/200][3/89]\tTotal Loss: 0.4447\tMAELoss: 0.2549\tMSSSIM Loss: 0.4809\n",
      "[0/200][4/89]\tTotal Loss: 0.2412\tMAELoss: 0.1518\tMSSSIM Loss: 0.2582\n",
      "[0/200][5/89]\tTotal Loss: 0.3501\tMAELoss: 0.3018\tMSSSIM Loss: 0.3593\n",
      "[0/200][6/89]\tTotal Loss: 0.3181\tMAELoss: 0.3144\tMSSSIM Loss: 0.3188\n",
      "[0/200][7/89]\tTotal Loss: 0.2617\tMAELoss: 0.2677\tMSSSIM Loss: 0.2605\n",
      "[0/200][8/89]\tTotal Loss: 0.2069\tMAELoss: 0.2260\tMSSSIM Loss: 0.2033\n",
      "[0/200][9/89]\tTotal Loss: 0.2308\tMAELoss: 0.1787\tMSSSIM Loss: 0.2407\n",
      "[0/200][10/89]\tTotal Loss: 0.1317\tMAELoss: 0.2021\tMSSSIM Loss: 0.1183\n",
      "[0/200][11/89]\tTotal Loss: 0.1967\tMAELoss: 0.1226\tMSSSIM Loss: 0.2108\n",
      "[0/200][12/89]\tTotal Loss: 0.1616\tMAELoss: 0.0864\tMSSSIM Loss: 0.1759\n",
      "[0/200][13/89]\tTotal Loss: 0.1757\tMAELoss: 0.1226\tMSSSIM Loss: 0.1858\n",
      "[0/200][14/89]\tTotal Loss: 0.1178\tMAELoss: 0.0824\tMSSSIM Loss: 0.1246\n",
      "[0/200][15/89]\tTotal Loss: 0.0684\tMAELoss: 0.0644\tMSSSIM Loss: 0.0692\n",
      "[0/200][16/89]\tTotal Loss: 0.0741\tMAELoss: 0.0513\tMSSSIM Loss: 0.0784\n",
      "[0/200][17/89]\tTotal Loss: 0.0731\tMAELoss: 0.0750\tMSSSIM Loss: 0.0727\n",
      "[0/200][18/89]\tTotal Loss: 0.0756\tMAELoss: 0.1055\tMSSSIM Loss: 0.0699\n",
      "[0/200][19/89]\tTotal Loss: 0.0619\tMAELoss: 0.0751\tMSSSIM Loss: 0.0594\n",
      "[0/200][20/89]\tTotal Loss: 0.0845\tMAELoss: 0.0733\tMSSSIM Loss: 0.0867\n",
      "[0/200][21/89]\tTotal Loss: 0.0633\tMAELoss: 0.0632\tMSSSIM Loss: 0.0633\n",
      "[0/200][22/89]\tTotal Loss: 0.0612\tMAELoss: 0.0491\tMSSSIM Loss: 0.0635\n",
      "[0/200][23/89]\tTotal Loss: 0.0486\tMAELoss: 0.0445\tMSSSIM Loss: 0.0494\n",
      "[0/200][24/89]\tTotal Loss: 0.0526\tMAELoss: 0.0606\tMSSSIM Loss: 0.0511\n",
      "[0/200][25/89]\tTotal Loss: 0.0569\tMAELoss: 0.0760\tMSSSIM Loss: 0.0532\n",
      "[0/200][26/89]\tTotal Loss: 0.0488\tMAELoss: 0.0639\tMSSSIM Loss: 0.0460\n",
      "[0/200][27/89]\tTotal Loss: 0.0536\tMAELoss: 0.0422\tMSSSIM Loss: 0.0558\n",
      "[0/200][28/89]\tTotal Loss: 0.0703\tMAELoss: 0.0526\tMSSSIM Loss: 0.0737\n",
      "[0/200][29/89]\tTotal Loss: 0.0345\tMAELoss: 0.0354\tMSSSIM Loss: 0.0343\n",
      "[0/200][30/89]\tTotal Loss: 0.0337\tMAELoss: 0.0400\tMSSSIM Loss: 0.0325\n",
      "[0/200][31/89]\tTotal Loss: 0.0397\tMAELoss: 0.0712\tMSSSIM Loss: 0.0337\n",
      "[0/200][32/89]\tTotal Loss: 0.0456\tMAELoss: 0.0597\tMSSSIM Loss: 0.0430\n",
      "[0/200][33/89]\tTotal Loss: 0.0492\tMAELoss: 0.0483\tMSSSIM Loss: 0.0493\n",
      "[0/200][34/89]\tTotal Loss: 0.0396\tMAELoss: 0.0340\tMSSSIM Loss: 0.0407\n",
      "[0/200][35/89]\tTotal Loss: 0.0341\tMAELoss: 0.0385\tMSSSIM Loss: 0.0333\n",
      "[0/200][36/89]\tTotal Loss: 0.0411\tMAELoss: 0.0362\tMSSSIM Loss: 0.0420\n",
      "[0/200][37/89]\tTotal Loss: 0.0371\tMAELoss: 0.0467\tMSSSIM Loss: 0.0353\n",
      "[0/200][38/89]\tTotal Loss: 0.0466\tMAELoss: 0.0572\tMSSSIM Loss: 0.0446\n",
      "[0/200][39/89]\tTotal Loss: 0.0450\tMAELoss: 0.0708\tMSSSIM Loss: 0.0400\n",
      "[0/200][40/89]\tTotal Loss: 0.0334\tMAELoss: 0.0542\tMSSSIM Loss: 0.0295\n",
      "[0/200][41/89]\tTotal Loss: 0.0301\tMAELoss: 0.0310\tMSSSIM Loss: 0.0299\n",
      "[0/200][42/89]\tTotal Loss: 0.0494\tMAELoss: 0.0515\tMSSSIM Loss: 0.0490\n",
      "[0/200][43/89]\tTotal Loss: 0.0369\tMAELoss: 0.0350\tMSSSIM Loss: 0.0372\n",
      "[0/200][44/89]\tTotal Loss: 0.0376\tMAELoss: 0.0377\tMSSSIM Loss: 0.0376\n",
      "[0/200][45/89]\tTotal Loss: 0.0357\tMAELoss: 0.0432\tMSSSIM Loss: 0.0343\n",
      "[0/200][46/89]\tTotal Loss: 0.0343\tMAELoss: 0.0451\tMSSSIM Loss: 0.0322\n",
      "[0/200][47/89]\tTotal Loss: 0.0458\tMAELoss: 0.0734\tMSSSIM Loss: 0.0405\n",
      "[0/200][48/89]\tTotal Loss: 0.0311\tMAELoss: 0.0400\tMSSSIM Loss: 0.0294\n",
      "[0/200][49/89]\tTotal Loss: 0.0358\tMAELoss: 0.0339\tMSSSIM Loss: 0.0361\n",
      "[0/200][50/89]\tTotal Loss: 0.0342\tMAELoss: 0.0316\tMSSSIM Loss: 0.0347\n",
      "[0/200][51/89]\tTotal Loss: 0.0479\tMAELoss: 0.0401\tMSSSIM Loss: 0.0493\n",
      "[0/200][52/89]\tTotal Loss: 0.0298\tMAELoss: 0.0243\tMSSSIM Loss: 0.0309\n",
      "[0/200][53/89]\tTotal Loss: 0.0295\tMAELoss: 0.0314\tMSSSIM Loss: 0.0292\n",
      "[0/200][54/89]\tTotal Loss: 0.0390\tMAELoss: 0.0561\tMSSSIM Loss: 0.0358\n",
      "[0/200][55/89]\tTotal Loss: 0.0433\tMAELoss: 0.0630\tMSSSIM Loss: 0.0396\n",
      "[0/200][56/89]\tTotal Loss: 0.0337\tMAELoss: 0.0482\tMSSSIM Loss: 0.0309\n",
      "[0/200][57/89]\tTotal Loss: 0.0276\tMAELoss: 0.0311\tMSSSIM Loss: 0.0270\n",
      "[0/200][58/89]\tTotal Loss: 0.0324\tMAELoss: 0.0323\tMSSSIM Loss: 0.0324\n",
      "[0/200][59/89]\tTotal Loss: 0.0437\tMAELoss: 0.0424\tMSSSIM Loss: 0.0439\n",
      "[0/200][60/89]\tTotal Loss: 0.0309\tMAELoss: 0.0367\tMSSSIM Loss: 0.0298\n",
      "[0/200][61/89]\tTotal Loss: 0.0249\tMAELoss: 0.0524\tMSSSIM Loss: 0.0196\n",
      "[0/200][62/89]\tTotal Loss: 0.0281\tMAELoss: 0.0383\tMSSSIM Loss: 0.0261\n",
      "[0/200][63/89]\tTotal Loss: 0.0318\tMAELoss: 0.0347\tMSSSIM Loss: 0.0313\n",
      "[0/200][64/89]\tTotal Loss: 0.0247\tMAELoss: 0.0257\tMSSSIM Loss: 0.0245\n",
      "[0/200][65/89]\tTotal Loss: 0.0234\tMAELoss: 0.0282\tMSSSIM Loss: 0.0224\n",
      "[0/200][66/89]\tTotal Loss: 0.0292\tMAELoss: 0.0259\tMSSSIM Loss: 0.0299\n",
      "[0/200][67/89]\tTotal Loss: 0.0385\tMAELoss: 0.0346\tMSSSIM Loss: 0.0392\n",
      "[0/200][68/89]\tTotal Loss: 0.0220\tMAELoss: 0.0228\tMSSSIM Loss: 0.0219\n",
      "[0/200][69/89]\tTotal Loss: 0.0245\tMAELoss: 0.0312\tMSSSIM Loss: 0.0233\n",
      "[0/200][70/89]\tTotal Loss: 0.0262\tMAELoss: 0.0240\tMSSSIM Loss: 0.0267\n",
      "[0/200][71/89]\tTotal Loss: 0.0282\tMAELoss: 0.0272\tMSSSIM Loss: 0.0284\n",
      "[0/200][72/89]\tTotal Loss: 0.0256\tMAELoss: 0.0305\tMSSSIM Loss: 0.0247\n",
      "[0/200][73/89]\tTotal Loss: 0.0296\tMAELoss: 0.0286\tMSSSIM Loss: 0.0298\n",
      "[0/200][74/89]\tTotal Loss: 0.0241\tMAELoss: 0.0246\tMSSSIM Loss: 0.0241\n",
      "[0/200][75/89]\tTotal Loss: 0.0215\tMAELoss: 0.0279\tMSSSIM Loss: 0.0203\n",
      "[0/200][76/89]\tTotal Loss: 0.0278\tMAELoss: 0.0333\tMSSSIM Loss: 0.0268\n",
      "[0/200][77/89]\tTotal Loss: 0.0269\tMAELoss: 0.0319\tMSSSIM Loss: 0.0259\n",
      "[0/200][78/89]\tTotal Loss: 0.0210\tMAELoss: 0.0288\tMSSSIM Loss: 0.0196\n",
      "[0/200][79/89]\tTotal Loss: 0.0205\tMAELoss: 0.0237\tMSSSIM Loss: 0.0198\n",
      "[0/200][80/89]\tTotal Loss: 0.0272\tMAELoss: 0.0271\tMSSSIM Loss: 0.0272\n",
      "[0/200][81/89]\tTotal Loss: 0.0157\tMAELoss: 0.0224\tMSSSIM Loss: 0.0144\n",
      "[0/200][82/89]\tTotal Loss: 0.0349\tMAELoss: 0.0345\tMSSSIM Loss: 0.0350\n",
      "[0/200][83/89]\tTotal Loss: 0.0229\tMAELoss: 0.0188\tMSSSIM Loss: 0.0237\n",
      "[0/200][84/89]\tTotal Loss: 0.0270\tMAELoss: 0.0245\tMSSSIM Loss: 0.0275\n",
      "[0/200][85/89]\tTotal Loss: 0.0384\tMAELoss: 0.0303\tMSSSIM Loss: 0.0399\n",
      "[0/200][86/89]\tTotal Loss: 0.0246\tMAELoss: 0.0242\tMSSSIM Loss: 0.0246\n",
      "[0/200][87/89]\tTotal Loss: 0.0264\tMAELoss: 0.0271\tMSSSIM Loss: 0.0263\n",
      "[0/200][88/89]\tTotal Loss: 0.0345\tMAELoss: 0.0393\tMSSSIM Loss: 0.0336\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    42: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    53: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    67: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    79: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    90: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch   110: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   121: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_7.tar\n",
      "Fold 8\n",
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.7116\tMAELoss: 0.2986\tMSSSIM Loss: 0.7902\n",
      "[0/200][1/89]\tTotal Loss: 0.6484\tMAELoss: 0.2969\tMSSSIM Loss: 0.7154\n",
      "[0/200][2/89]\tTotal Loss: 0.5162\tMAELoss: 0.2526\tMSSSIM Loss: 0.5665\n",
      "[0/200][3/89]\tTotal Loss: 0.3843\tMAELoss: 0.2604\tMSSSIM Loss: 0.4079\n",
      "[0/200][4/89]\tTotal Loss: 0.2870\tMAELoss: 0.1402\tMSSSIM Loss: 0.3149\n",
      "[0/200][5/89]\tTotal Loss: 0.2382\tMAELoss: 0.2782\tMSSSIM Loss: 0.2306\n",
      "[0/200][6/89]\tTotal Loss: 0.2684\tMAELoss: 0.2408\tMSSSIM Loss: 0.2736\n",
      "[0/200][7/89]\tTotal Loss: 0.2010\tMAELoss: 0.2711\tMSSSIM Loss: 0.1877\n",
      "[0/200][8/89]\tTotal Loss: 0.1884\tMAELoss: 0.2075\tMSSSIM Loss: 0.1848\n",
      "[0/200][9/89]\tTotal Loss: 0.1202\tMAELoss: 0.2607\tMSSSIM Loss: 0.0934\n",
      "[0/200][10/89]\tTotal Loss: 0.1164\tMAELoss: 0.1950\tMSSSIM Loss: 0.1014\n",
      "[0/200][11/89]\tTotal Loss: 0.1478\tMAELoss: 0.1352\tMSSSIM Loss: 0.1502\n",
      "[0/200][12/89]\tTotal Loss: 0.1131\tMAELoss: 0.0823\tMSSSIM Loss: 0.1189\n",
      "[0/200][13/89]\tTotal Loss: 0.0835\tMAELoss: 0.0676\tMSSSIM Loss: 0.0865\n",
      "[0/200][14/89]\tTotal Loss: 0.0866\tMAELoss: 0.1080\tMSSSIM Loss: 0.0825\n",
      "[0/200][15/89]\tTotal Loss: 0.0644\tMAELoss: 0.0985\tMSSSIM Loss: 0.0579\n",
      "[0/200][16/89]\tTotal Loss: 0.0449\tMAELoss: 0.0574\tMSSSIM Loss: 0.0426\n",
      "[0/200][17/89]\tTotal Loss: 0.0437\tMAELoss: 0.0448\tMSSSIM Loss: 0.0435\n",
      "[0/200][18/89]\tTotal Loss: 0.0839\tMAELoss: 0.0639\tMSSSIM Loss: 0.0877\n",
      "[0/200][19/89]\tTotal Loss: 0.0385\tMAELoss: 0.0360\tMSSSIM Loss: 0.0389\n",
      "[0/200][20/89]\tTotal Loss: 0.0400\tMAELoss: 0.0528\tMSSSIM Loss: 0.0376\n",
      "[0/200][21/89]\tTotal Loss: 0.0530\tMAELoss: 0.0923\tMSSSIM Loss: 0.0455\n",
      "[0/200][22/89]\tTotal Loss: 0.0558\tMAELoss: 0.1089\tMSSSIM Loss: 0.0456\n",
      "[0/200][23/89]\tTotal Loss: 0.0330\tMAELoss: 0.0517\tMSSSIM Loss: 0.0295\n",
      "[0/200][24/89]\tTotal Loss: 0.0387\tMAELoss: 0.0425\tMSSSIM Loss: 0.0380\n",
      "[0/200][25/89]\tTotal Loss: 0.0675\tMAELoss: 0.0581\tMSSSIM Loss: 0.0693\n",
      "[0/200][26/89]\tTotal Loss: 0.0520\tMAELoss: 0.0486\tMSSSIM Loss: 0.0526\n",
      "[0/200][27/89]\tTotal Loss: 0.0417\tMAELoss: 0.0359\tMSSSIM Loss: 0.0428\n",
      "[0/200][28/89]\tTotal Loss: 0.0649\tMAELoss: 0.0702\tMSSSIM Loss: 0.0639\n",
      "[0/200][29/89]\tTotal Loss: 0.0555\tMAELoss: 0.0793\tMSSSIM Loss: 0.0509\n",
      "[0/200][30/89]\tTotal Loss: 0.0332\tMAELoss: 0.0512\tMSSSIM Loss: 0.0298\n",
      "[0/200][31/89]\tTotal Loss: 0.0441\tMAELoss: 0.0729\tMSSSIM Loss: 0.0386\n",
      "[0/200][32/89]\tTotal Loss: 0.0536\tMAELoss: 0.0536\tMSSSIM Loss: 0.0535\n",
      "[0/200][33/89]\tTotal Loss: 0.0460\tMAELoss: 0.0534\tMSSSIM Loss: 0.0446\n",
      "[0/200][34/89]\tTotal Loss: 0.0527\tMAELoss: 0.0467\tMSSSIM Loss: 0.0539\n",
      "[0/200][35/89]\tTotal Loss: 0.0414\tMAELoss: 0.0415\tMSSSIM Loss: 0.0414\n",
      "[0/200][36/89]\tTotal Loss: 0.0323\tMAELoss: 0.0251\tMSSSIM Loss: 0.0337\n",
      "[0/200][37/89]\tTotal Loss: 0.0388\tMAELoss: 0.0475\tMSSSIM Loss: 0.0372\n",
      "[0/200][38/89]\tTotal Loss: 0.0397\tMAELoss: 0.0522\tMSSSIM Loss: 0.0374\n",
      "[0/200][39/89]\tTotal Loss: 0.0397\tMAELoss: 0.0558\tMSSSIM Loss: 0.0366\n",
      "[0/200][40/89]\tTotal Loss: 0.0344\tMAELoss: 0.0480\tMSSSIM Loss: 0.0318\n",
      "[0/200][41/89]\tTotal Loss: 0.0399\tMAELoss: 0.0336\tMSSSIM Loss: 0.0411\n",
      "[0/200][42/89]\tTotal Loss: 0.0342\tMAELoss: 0.0318\tMSSSIM Loss: 0.0347\n",
      "[0/200][43/89]\tTotal Loss: 0.0307\tMAELoss: 0.0423\tMSSSIM Loss: 0.0285\n",
      "[0/200][44/89]\tTotal Loss: 0.0343\tMAELoss: 0.0539\tMSSSIM Loss: 0.0306\n",
      "[0/200][45/89]\tTotal Loss: 0.0440\tMAELoss: 0.0500\tMSSSIM Loss: 0.0428\n",
      "[0/200][46/89]\tTotal Loss: 0.0281\tMAELoss: 0.0392\tMSSSIM Loss: 0.0259\n",
      "[0/200][47/89]\tTotal Loss: 0.0295\tMAELoss: 0.0311\tMSSSIM Loss: 0.0292\n",
      "[0/200][48/89]\tTotal Loss: 0.0331\tMAELoss: 0.0330\tMSSSIM Loss: 0.0332\n",
      "[0/200][49/89]\tTotal Loss: 0.0265\tMAELoss: 0.0303\tMSSSIM Loss: 0.0258\n",
      "[0/200][50/89]\tTotal Loss: 0.0414\tMAELoss: 0.0469\tMSSSIM Loss: 0.0404\n",
      "[0/200][51/89]\tTotal Loss: 0.0383\tMAELoss: 0.0429\tMSSSIM Loss: 0.0374\n",
      "[0/200][52/89]\tTotal Loss: 0.0282\tMAELoss: 0.0282\tMSSSIM Loss: 0.0282\n",
      "[0/200][53/89]\tTotal Loss: 0.0303\tMAELoss: 0.0300\tMSSSIM Loss: 0.0304\n",
      "[0/200][54/89]\tTotal Loss: 0.0316\tMAELoss: 0.0332\tMSSSIM Loss: 0.0313\n",
      "[0/200][55/89]\tTotal Loss: 0.0275\tMAELoss: 0.0240\tMSSSIM Loss: 0.0282\n",
      "[0/200][56/89]\tTotal Loss: 0.0389\tMAELoss: 0.0451\tMSSSIM Loss: 0.0377\n",
      "[0/200][57/89]\tTotal Loss: 0.0256\tMAELoss: 0.0415\tMSSSIM Loss: 0.0225\n",
      "[0/200][58/89]\tTotal Loss: 0.0345\tMAELoss: 0.0582\tMSSSIM Loss: 0.0300\n",
      "[0/200][59/89]\tTotal Loss: 0.0276\tMAELoss: 0.0385\tMSSSIM Loss: 0.0255\n",
      "[0/200][60/89]\tTotal Loss: 0.0316\tMAELoss: 0.0340\tMSSSIM Loss: 0.0312\n",
      "[0/200][61/89]\tTotal Loss: 0.0496\tMAELoss: 0.0410\tMSSSIM Loss: 0.0512\n",
      "[0/200][62/89]\tTotal Loss: 0.0438\tMAELoss: 0.0366\tMSSSIM Loss: 0.0451\n",
      "[0/200][63/89]\tTotal Loss: 0.0373\tMAELoss: 0.0329\tMSSSIM Loss: 0.0381\n",
      "[0/200][64/89]\tTotal Loss: 0.0370\tMAELoss: 0.0523\tMSSSIM Loss: 0.0341\n",
      "[0/200][65/89]\tTotal Loss: 0.0344\tMAELoss: 0.0540\tMSSSIM Loss: 0.0306\n",
      "[0/200][66/89]\tTotal Loss: 0.0376\tMAELoss: 0.0459\tMSSSIM Loss: 0.0360\n",
      "[0/200][67/89]\tTotal Loss: 0.0241\tMAELoss: 0.0301\tMSSSIM Loss: 0.0229\n",
      "[0/200][68/89]\tTotal Loss: 0.0449\tMAELoss: 0.0378\tMSSSIM Loss: 0.0463\n",
      "[0/200][69/89]\tTotal Loss: 0.0357\tMAELoss: 0.0286\tMSSSIM Loss: 0.0371\n",
      "[0/200][70/89]\tTotal Loss: 0.0211\tMAELoss: 0.0359\tMSSSIM Loss: 0.0182\n",
      "[0/200][71/89]\tTotal Loss: 0.0359\tMAELoss: 0.0578\tMSSSIM Loss: 0.0317\n",
      "[0/200][72/89]\tTotal Loss: 0.0425\tMAELoss: 0.0578\tMSSSIM Loss: 0.0395\n",
      "[0/200][73/89]\tTotal Loss: 0.0410\tMAELoss: 0.0543\tMSSSIM Loss: 0.0384\n",
      "[0/200][74/89]\tTotal Loss: 0.0286\tMAELoss: 0.0261\tMSSSIM Loss: 0.0290\n",
      "[0/200][75/89]\tTotal Loss: 0.0319\tMAELoss: 0.0289\tMSSSIM Loss: 0.0325\n",
      "[0/200][76/89]\tTotal Loss: 0.0398\tMAELoss: 0.0343\tMSSSIM Loss: 0.0408\n",
      "[0/200][77/89]\tTotal Loss: 0.0246\tMAELoss: 0.0286\tMSSSIM Loss: 0.0239\n",
      "[0/200][78/89]\tTotal Loss: 0.0316\tMAELoss: 0.0407\tMSSSIM Loss: 0.0299\n",
      "[0/200][79/89]\tTotal Loss: 0.0268\tMAELoss: 0.0382\tMSSSIM Loss: 0.0246\n",
      "[0/200][80/89]\tTotal Loss: 0.0355\tMAELoss: 0.0548\tMSSSIM Loss: 0.0318\n",
      "[0/200][81/89]\tTotal Loss: 0.0249\tMAELoss: 0.0301\tMSSSIM Loss: 0.0239\n",
      "[0/200][82/89]\tTotal Loss: 0.0327\tMAELoss: 0.0343\tMSSSIM Loss: 0.0324\n",
      "[0/200][83/89]\tTotal Loss: 0.0258\tMAELoss: 0.0329\tMSSSIM Loss: 0.0244\n",
      "[0/200][84/89]\tTotal Loss: 0.0424\tMAELoss: 0.0327\tMSSSIM Loss: 0.0443\n",
      "[0/200][85/89]\tTotal Loss: 0.0294\tMAELoss: 0.0358\tMSSSIM Loss: 0.0282\n",
      "[0/200][86/89]\tTotal Loss: 0.0219\tMAELoss: 0.0226\tMSSSIM Loss: 0.0217\n",
      "[0/200][87/89]\tTotal Loss: 0.0375\tMAELoss: 0.0323\tMSSSIM Loss: 0.0385\n",
      "[0/200][88/89]\tTotal Loss: 0.0307\tMAELoss: 0.0378\tMSSSIM Loss: 0.0293\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    29: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    40: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    51: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    72: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    83: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch    94: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   105: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_8.tar\n",
      "Fold 9\n",
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.6944\tMAELoss: 0.3081\tMSSSIM Loss: 0.7680\n",
      "[0/200][1/89]\tTotal Loss: 0.7015\tMAELoss: 0.2875\tMSSSIM Loss: 0.7804\n",
      "[0/200][2/89]\tTotal Loss: 0.6199\tMAELoss: 0.3063\tMSSSIM Loss: 0.6796\n",
      "[0/200][3/89]\tTotal Loss: 0.4040\tMAELoss: 0.2780\tMSSSIM Loss: 0.4280\n",
      "[0/200][4/89]\tTotal Loss: 0.2436\tMAELoss: 0.2640\tMSSSIM Loss: 0.2397\n",
      "[0/200][5/89]\tTotal Loss: 0.2620\tMAELoss: 0.4003\tMSSSIM Loss: 0.2357\n",
      "[0/200][6/89]\tTotal Loss: 0.2597\tMAELoss: 0.4156\tMSSSIM Loss: 0.2300\n",
      "[0/200][7/89]\tTotal Loss: 0.2236\tMAELoss: 0.3851\tMSSSIM Loss: 0.1929\n",
      "[0/200][8/89]\tTotal Loss: 0.1956\tMAELoss: 0.2602\tMSSSIM Loss: 0.1833\n",
      "[0/200][9/89]\tTotal Loss: 0.2205\tMAELoss: 0.1781\tMSSSIM Loss: 0.2286\n",
      "[0/200][10/89]\tTotal Loss: 0.2245\tMAELoss: 0.1441\tMSSSIM Loss: 0.2399\n",
      "[0/200][11/89]\tTotal Loss: 0.1479\tMAELoss: 0.1305\tMSSSIM Loss: 0.1513\n",
      "[0/200][12/89]\tTotal Loss: 0.1375\tMAELoss: 0.0735\tMSSSIM Loss: 0.1497\n",
      "[0/200][13/89]\tTotal Loss: 0.1809\tMAELoss: 0.0919\tMSSSIM Loss: 0.1978\n",
      "[0/200][14/89]\tTotal Loss: 0.1129\tMAELoss: 0.0789\tMSSSIM Loss: 0.1194\n",
      "[0/200][15/89]\tTotal Loss: 0.1108\tMAELoss: 0.0776\tMSSSIM Loss: 0.1171\n",
      "[0/200][16/89]\tTotal Loss: 0.1674\tMAELoss: 0.1328\tMSSSIM Loss: 0.1740\n",
      "[0/200][17/89]\tTotal Loss: 0.0647\tMAELoss: 0.1021\tMSSSIM Loss: 0.0575\n",
      "[0/200][18/89]\tTotal Loss: 0.0456\tMAELoss: 0.1028\tMSSSIM Loss: 0.0346\n",
      "[0/200][19/89]\tTotal Loss: 0.0458\tMAELoss: 0.0617\tMSSSIM Loss: 0.0428\n",
      "[0/200][20/89]\tTotal Loss: 0.0469\tMAELoss: 0.0546\tMSSSIM Loss: 0.0455\n",
      "[0/200][21/89]\tTotal Loss: 0.0531\tMAELoss: 0.0580\tMSSSIM Loss: 0.0521\n",
      "[0/200][22/89]\tTotal Loss: 0.0685\tMAELoss: 0.0728\tMSSSIM Loss: 0.0676\n",
      "[0/200][23/89]\tTotal Loss: 0.0559\tMAELoss: 0.0581\tMSSSIM Loss: 0.0555\n",
      "[0/200][24/89]\tTotal Loss: 0.0517\tMAELoss: 0.0554\tMSSSIM Loss: 0.0510\n",
      "[0/200][25/89]\tTotal Loss: 0.0347\tMAELoss: 0.0372\tMSSSIM Loss: 0.0342\n",
      "[0/200][26/89]\tTotal Loss: 0.0335\tMAELoss: 0.0336\tMSSSIM Loss: 0.0335\n",
      "[0/200][27/89]\tTotal Loss: 0.0347\tMAELoss: 0.0316\tMSSSIM Loss: 0.0353\n",
      "[0/200][28/89]\tTotal Loss: 0.0467\tMAELoss: 0.0335\tMSSSIM Loss: 0.0492\n",
      "[0/200][29/89]\tTotal Loss: 0.0377\tMAELoss: 0.0581\tMSSSIM Loss: 0.0338\n",
      "[0/200][30/89]\tTotal Loss: 0.0522\tMAELoss: 0.0695\tMSSSIM Loss: 0.0489\n",
      "[0/200][31/89]\tTotal Loss: 0.0426\tMAELoss: 0.0802\tMSSSIM Loss: 0.0354\n",
      "[0/200][32/89]\tTotal Loss: 0.0406\tMAELoss: 0.0545\tMSSSIM Loss: 0.0380\n",
      "[0/200][33/89]\tTotal Loss: 0.0344\tMAELoss: 0.0368\tMSSSIM Loss: 0.0340\n",
      "[0/200][34/89]\tTotal Loss: 0.0328\tMAELoss: 0.0324\tMSSSIM Loss: 0.0329\n",
      "[0/200][35/89]\tTotal Loss: 0.0395\tMAELoss: 0.0407\tMSSSIM Loss: 0.0393\n",
      "[0/200][36/89]\tTotal Loss: 0.0405\tMAELoss: 0.0461\tMSSSIM Loss: 0.0395\n",
      "[0/200][37/89]\tTotal Loss: 0.0311\tMAELoss: 0.0424\tMSSSIM Loss: 0.0289\n",
      "[0/200][38/89]\tTotal Loss: 0.0301\tMAELoss: 0.0324\tMSSSIM Loss: 0.0296\n",
      "[0/200][39/89]\tTotal Loss: 0.0324\tMAELoss: 0.0313\tMSSSIM Loss: 0.0326\n",
      "[0/200][40/89]\tTotal Loss: 0.0382\tMAELoss: 0.0318\tMSSSIM Loss: 0.0394\n",
      "[0/200][41/89]\tTotal Loss: 0.0345\tMAELoss: 0.0308\tMSSSIM Loss: 0.0352\n",
      "[0/200][42/89]\tTotal Loss: 0.0305\tMAELoss: 0.0380\tMSSSIM Loss: 0.0290\n",
      "[0/200][43/89]\tTotal Loss: 0.0254\tMAELoss: 0.0359\tMSSSIM Loss: 0.0233\n",
      "[0/200][44/89]\tTotal Loss: 0.0396\tMAELoss: 0.0332\tMSSSIM Loss: 0.0409\n",
      "[0/200][45/89]\tTotal Loss: 0.0420\tMAELoss: 0.0540\tMSSSIM Loss: 0.0397\n",
      "[0/200][46/89]\tTotal Loss: 0.0207\tMAELoss: 0.0252\tMSSSIM Loss: 0.0198\n",
      "[0/200][47/89]\tTotal Loss: 0.0422\tMAELoss: 0.0659\tMSSSIM Loss: 0.0377\n",
      "[0/200][48/89]\tTotal Loss: 0.0358\tMAELoss: 0.0544\tMSSSIM Loss: 0.0323\n",
      "[0/200][49/89]\tTotal Loss: 0.0337\tMAELoss: 0.0312\tMSSSIM Loss: 0.0342\n",
      "[0/200][50/89]\tTotal Loss: 0.0301\tMAELoss: 0.0300\tMSSSIM Loss: 0.0301\n",
      "[0/200][51/89]\tTotal Loss: 0.0383\tMAELoss: 0.0387\tMSSSIM Loss: 0.0382\n",
      "[0/200][52/89]\tTotal Loss: 0.0386\tMAELoss: 0.0565\tMSSSIM Loss: 0.0352\n",
      "[0/200][53/89]\tTotal Loss: 0.0300\tMAELoss: 0.0499\tMSSSIM Loss: 0.0262\n",
      "[0/200][54/89]\tTotal Loss: 0.0523\tMAELoss: 0.0691\tMSSSIM Loss: 0.0491\n",
      "[0/200][55/89]\tTotal Loss: 0.0396\tMAELoss: 0.0523\tMSSSIM Loss: 0.0371\n",
      "[0/200][56/89]\tTotal Loss: 0.0357\tMAELoss: 0.0401\tMSSSIM Loss: 0.0348\n",
      "[0/200][57/89]\tTotal Loss: 0.0404\tMAELoss: 0.0371\tMSSSIM Loss: 0.0411\n",
      "[0/200][58/89]\tTotal Loss: 0.0410\tMAELoss: 0.0424\tMSSSIM Loss: 0.0407\n",
      "[0/200][59/89]\tTotal Loss: 0.0349\tMAELoss: 0.0352\tMSSSIM Loss: 0.0348\n",
      "[0/200][60/89]\tTotal Loss: 0.0342\tMAELoss: 0.0393\tMSSSIM Loss: 0.0332\n",
      "[0/200][61/89]\tTotal Loss: 0.0282\tMAELoss: 0.0517\tMSSSIM Loss: 0.0237\n",
      "[0/200][62/89]\tTotal Loss: 0.0391\tMAELoss: 0.0450\tMSSSIM Loss: 0.0380\n",
      "[0/200][63/89]\tTotal Loss: 0.0348\tMAELoss: 0.0408\tMSSSIM Loss: 0.0337\n",
      "[0/200][64/89]\tTotal Loss: 0.0339\tMAELoss: 0.0330\tMSSSIM Loss: 0.0340\n",
      "[0/200][65/89]\tTotal Loss: 0.0309\tMAELoss: 0.0335\tMSSSIM Loss: 0.0304\n",
      "[0/200][66/89]\tTotal Loss: 0.0341\tMAELoss: 0.0339\tMSSSIM Loss: 0.0341\n",
      "[0/200][67/89]\tTotal Loss: 0.0314\tMAELoss: 0.0358\tMSSSIM Loss: 0.0306\n",
      "[0/200][68/89]\tTotal Loss: 0.0280\tMAELoss: 0.0430\tMSSSIM Loss: 0.0252\n",
      "[0/200][69/89]\tTotal Loss: 0.0268\tMAELoss: 0.0378\tMSSSIM Loss: 0.0247\n",
      "[0/200][70/89]\tTotal Loss: 0.0315\tMAELoss: 0.0271\tMSSSIM Loss: 0.0323\n",
      "[0/200][71/89]\tTotal Loss: 0.0259\tMAELoss: 0.0304\tMSSSIM Loss: 0.0250\n",
      "[0/200][72/89]\tTotal Loss: 0.0273\tMAELoss: 0.0291\tMSSSIM Loss: 0.0269\n",
      "[0/200][73/89]\tTotal Loss: 0.0288\tMAELoss: 0.0295\tMSSSIM Loss: 0.0286\n",
      "[0/200][74/89]\tTotal Loss: 0.0329\tMAELoss: 0.0378\tMSSSIM Loss: 0.0320\n",
      "[0/200][75/89]\tTotal Loss: 0.0283\tMAELoss: 0.0353\tMSSSIM Loss: 0.0270\n",
      "[0/200][76/89]\tTotal Loss: 0.0340\tMAELoss: 0.0342\tMSSSIM Loss: 0.0339\n",
      "[0/200][77/89]\tTotal Loss: 0.0436\tMAELoss: 0.0286\tMSSSIM Loss: 0.0464\n",
      "[0/200][78/89]\tTotal Loss: 0.0292\tMAELoss: 0.0383\tMSSSIM Loss: 0.0275\n",
      "[0/200][79/89]\tTotal Loss: 0.0357\tMAELoss: 0.0477\tMSSSIM Loss: 0.0334\n",
      "[0/200][80/89]\tTotal Loss: 0.0309\tMAELoss: 0.0364\tMSSSIM Loss: 0.0298\n",
      "[0/200][81/89]\tTotal Loss: 0.0188\tMAELoss: 0.0287\tMSSSIM Loss: 0.0169\n",
      "[0/200][82/89]\tTotal Loss: 0.0269\tMAELoss: 0.0242\tMSSSIM Loss: 0.0274\n",
      "[0/200][83/89]\tTotal Loss: 0.0346\tMAELoss: 0.0385\tMSSSIM Loss: 0.0339\n",
      "[0/200][84/89]\tTotal Loss: 0.0379\tMAELoss: 0.0315\tMSSSIM Loss: 0.0391\n",
      "[0/200][85/89]\tTotal Loss: 0.0322\tMAELoss: 0.0382\tMSSSIM Loss: 0.0311\n",
      "[0/200][86/89]\tTotal Loss: 0.0264\tMAELoss: 0.0377\tMSSSIM Loss: 0.0242\n",
      "[0/200][87/89]\tTotal Loss: 0.0300\tMAELoss: 0.0377\tMSSSIM Loss: 0.0286\n",
      "[0/200][88/89]\tTotal Loss: 0.0293\tMAELoss: 0.0337\tMSSSIM Loss: 0.0284\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    19: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    44: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    56: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    67: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    78: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch    89: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   102: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_9.tar\n",
      "Fold 10\n",
      "FLAG: NO CHECKPOINT LOADED.\n",
      "0.001\n",
      "[0/200][0/89]\tTotal Loss: 0.7631\tMAELoss: 0.3078\tMSSSIM Loss: 0.8498\n",
      "[0/200][1/89]\tTotal Loss: 0.6496\tMAELoss: 0.3026\tMSSSIM Loss: 0.7156\n",
      "[0/200][2/89]\tTotal Loss: 0.5308\tMAELoss: 0.2932\tMSSSIM Loss: 0.5761\n",
      "[0/200][3/89]\tTotal Loss: 0.3929\tMAELoss: 0.2039\tMSSSIM Loss: 0.4289\n",
      "[0/200][4/89]\tTotal Loss: 0.2344\tMAELoss: 0.2950\tMSSSIM Loss: 0.2229\n",
      "[0/200][5/89]\tTotal Loss: 0.4029\tMAELoss: 0.3304\tMSSSIM Loss: 0.4167\n",
      "[0/200][6/89]\tTotal Loss: 0.3183\tMAELoss: 0.3964\tMSSSIM Loss: 0.3034\n",
      "[0/200][7/89]\tTotal Loss: 0.2822\tMAELoss: 0.3046\tMSSSIM Loss: 0.2779\n",
      "[0/200][8/89]\tTotal Loss: 0.1864\tMAELoss: 0.2409\tMSSSIM Loss: 0.1760\n",
      "[0/200][9/89]\tTotal Loss: 0.2131\tMAELoss: 0.1806\tMSSSIM Loss: 0.2193\n",
      "[0/200][10/89]\tTotal Loss: 0.1863\tMAELoss: 0.1955\tMSSSIM Loss: 0.1845\n",
      "[0/200][11/89]\tTotal Loss: 0.1499\tMAELoss: 0.1279\tMSSSIM Loss: 0.1541\n",
      "[0/200][12/89]\tTotal Loss: 0.1236\tMAELoss: 0.0979\tMSSSIM Loss: 0.1284\n",
      "[0/200][13/89]\tTotal Loss: 0.1595\tMAELoss: 0.0956\tMSSSIM Loss: 0.1716\n",
      "[0/200][14/89]\tTotal Loss: 0.2758\tMAELoss: 0.1408\tMSSSIM Loss: 0.3015\n",
      "[0/200][15/89]\tTotal Loss: 0.1155\tMAELoss: 0.0700\tMSSSIM Loss: 0.1242\n",
      "[0/200][16/89]\tTotal Loss: 0.1521\tMAELoss: 0.0889\tMSSSIM Loss: 0.1641\n",
      "[0/200][17/89]\tTotal Loss: 0.1373\tMAELoss: 0.0940\tMSSSIM Loss: 0.1455\n",
      "[0/200][18/89]\tTotal Loss: 0.0941\tMAELoss: 0.0753\tMSSSIM Loss: 0.0977\n",
      "[0/200][19/89]\tTotal Loss: 0.0627\tMAELoss: 0.0565\tMSSSIM Loss: 0.0639\n",
      "[0/200][20/89]\tTotal Loss: 0.0487\tMAELoss: 0.0768\tMSSSIM Loss: 0.0434\n",
      "[0/200][21/89]\tTotal Loss: 0.0772\tMAELoss: 0.1125\tMSSSIM Loss: 0.0705\n",
      "[0/200][22/89]\tTotal Loss: 0.0810\tMAELoss: 0.1063\tMSSSIM Loss: 0.0762\n",
      "[0/200][23/89]\tTotal Loss: 0.0513\tMAELoss: 0.0706\tMSSSIM Loss: 0.0477\n",
      "[0/200][24/89]\tTotal Loss: 0.0475\tMAELoss: 0.0540\tMSSSIM Loss: 0.0462\n",
      "[0/200][25/89]\tTotal Loss: 0.0663\tMAELoss: 0.0583\tMSSSIM Loss: 0.0678\n",
      "[0/200][26/89]\tTotal Loss: 0.0447\tMAELoss: 0.0361\tMSSSIM Loss: 0.0464\n",
      "[0/200][27/89]\tTotal Loss: 0.0405\tMAELoss: 0.0527\tMSSSIM Loss: 0.0382\n",
      "[0/200][28/89]\tTotal Loss: 0.0470\tMAELoss: 0.0835\tMSSSIM Loss: 0.0401\n",
      "[0/200][29/89]\tTotal Loss: 0.0666\tMAELoss: 0.1072\tMSSSIM Loss: 0.0588\n",
      "[0/200][30/89]\tTotal Loss: 0.0422\tMAELoss: 0.0812\tMSSSIM Loss: 0.0348\n",
      "[0/200][31/89]\tTotal Loss: 0.0365\tMAELoss: 0.0490\tMSSSIM Loss: 0.0341\n",
      "[0/200][32/89]\tTotal Loss: 0.0388\tMAELoss: 0.0362\tMSSSIM Loss: 0.0393\n",
      "[0/200][33/89]\tTotal Loss: 0.0479\tMAELoss: 0.0454\tMSSSIM Loss: 0.0483\n",
      "[0/200][34/89]\tTotal Loss: 0.0356\tMAELoss: 0.0321\tMSSSIM Loss: 0.0363\n",
      "[0/200][35/89]\tTotal Loss: 0.0337\tMAELoss: 0.0379\tMSSSIM Loss: 0.0329\n",
      "[0/200][36/89]\tTotal Loss: 0.0418\tMAELoss: 0.0451\tMSSSIM Loss: 0.0412\n",
      "[0/200][37/89]\tTotal Loss: 0.0320\tMAELoss: 0.0431\tMSSSIM Loss: 0.0299\n",
      "[0/200][38/89]\tTotal Loss: 0.0444\tMAELoss: 0.0539\tMSSSIM Loss: 0.0425\n",
      "[0/200][39/89]\tTotal Loss: 0.0301\tMAELoss: 0.0390\tMSSSIM Loss: 0.0284\n",
      "[0/200][40/89]\tTotal Loss: 0.0318\tMAELoss: 0.0369\tMSSSIM Loss: 0.0308\n",
      "[0/200][41/89]\tTotal Loss: 0.0289\tMAELoss: 0.0341\tMSSSIM Loss: 0.0279\n",
      "[0/200][42/89]\tTotal Loss: 0.0349\tMAELoss: 0.0331\tMSSSIM Loss: 0.0352\n",
      "[0/200][43/89]\tTotal Loss: 0.0533\tMAELoss: 0.0481\tMSSSIM Loss: 0.0543\n",
      "[0/200][44/89]\tTotal Loss: 0.0321\tMAELoss: 0.0291\tMSSSIM Loss: 0.0326\n",
      "[0/200][45/89]\tTotal Loss: 0.0356\tMAELoss: 0.0369\tMSSSIM Loss: 0.0354\n",
      "[0/200][46/89]\tTotal Loss: 0.0330\tMAELoss: 0.0510\tMSSSIM Loss: 0.0295\n",
      "[0/200][47/89]\tTotal Loss: 0.0358\tMAELoss: 0.0617\tMSSSIM Loss: 0.0309\n",
      "[0/200][48/89]\tTotal Loss: 0.0266\tMAELoss: 0.0521\tMSSSIM Loss: 0.0218\n",
      "[0/200][49/89]\tTotal Loss: 0.0465\tMAELoss: 0.0555\tMSSSIM Loss: 0.0447\n",
      "[0/200][50/89]\tTotal Loss: 0.0287\tMAELoss: 0.0289\tMSSSIM Loss: 0.0287\n",
      "[0/200][51/89]\tTotal Loss: 0.0250\tMAELoss: 0.0283\tMSSSIM Loss: 0.0243\n",
      "[0/200][52/89]\tTotal Loss: 0.0349\tMAELoss: 0.0220\tMSSSIM Loss: 0.0374\n",
      "[0/200][53/89]\tTotal Loss: 0.0517\tMAELoss: 0.0419\tMSSSIM Loss: 0.0536\n",
      "[0/200][54/89]\tTotal Loss: 0.0247\tMAELoss: 0.0269\tMSSSIM Loss: 0.0243\n",
      "[0/200][55/89]\tTotal Loss: 0.0288\tMAELoss: 0.0278\tMSSSIM Loss: 0.0290\n",
      "[0/200][56/89]\tTotal Loss: 0.0257\tMAELoss: 0.0272\tMSSSIM Loss: 0.0254\n",
      "[0/200][57/89]\tTotal Loss: 0.0345\tMAELoss: 0.0346\tMSSSIM Loss: 0.0345\n",
      "[0/200][58/89]\tTotal Loss: 0.0250\tMAELoss: 0.0365\tMSSSIM Loss: 0.0229\n",
      "[0/200][59/89]\tTotal Loss: 0.0333\tMAELoss: 0.0313\tMSSSIM Loss: 0.0337\n",
      "[0/200][60/89]\tTotal Loss: 0.0248\tMAELoss: 0.0305\tMSSSIM Loss: 0.0237\n",
      "[0/200][61/89]\tTotal Loss: 0.0278\tMAELoss: 0.0345\tMSSSIM Loss: 0.0265\n",
      "[0/200][62/89]\tTotal Loss: 0.0185\tMAELoss: 0.0171\tMSSSIM Loss: 0.0187\n",
      "[0/200][63/89]\tTotal Loss: 0.0361\tMAELoss: 0.0338\tMSSSIM Loss: 0.0366\n",
      "[0/200][64/89]\tTotal Loss: 0.0221\tMAELoss: 0.0329\tMSSSIM Loss: 0.0200\n",
      "[0/200][65/89]\tTotal Loss: 0.0352\tMAELoss: 0.0563\tMSSSIM Loss: 0.0312\n",
      "[0/200][66/89]\tTotal Loss: 0.0333\tMAELoss: 0.0421\tMSSSIM Loss: 0.0316\n",
      "[0/200][67/89]\tTotal Loss: 0.0305\tMAELoss: 0.0328\tMSSSIM Loss: 0.0300\n",
      "[0/200][68/89]\tTotal Loss: 0.0291\tMAELoss: 0.0304\tMSSSIM Loss: 0.0289\n",
      "[0/200][69/89]\tTotal Loss: 0.0326\tMAELoss: 0.0338\tMSSSIM Loss: 0.0323\n",
      "[0/200][70/89]\tTotal Loss: 0.0238\tMAELoss: 0.0280\tMSSSIM Loss: 0.0230\n",
      "[0/200][71/89]\tTotal Loss: 0.0320\tMAELoss: 0.0329\tMSSSIM Loss: 0.0319\n",
      "[0/200][72/89]\tTotal Loss: 0.0299\tMAELoss: 0.0339\tMSSSIM Loss: 0.0292\n",
      "[0/200][73/89]\tTotal Loss: 0.0325\tMAELoss: 0.0335\tMSSSIM Loss: 0.0323\n",
      "[0/200][74/89]\tTotal Loss: 0.0455\tMAELoss: 0.0378\tMSSSIM Loss: 0.0470\n",
      "[0/200][75/89]\tTotal Loss: 0.0265\tMAELoss: 0.0300\tMSSSIM Loss: 0.0258\n",
      "[0/200][76/89]\tTotal Loss: 0.0384\tMAELoss: 0.0360\tMSSSIM Loss: 0.0389\n",
      "[0/200][77/89]\tTotal Loss: 0.0265\tMAELoss: 0.0202\tMSSSIM Loss: 0.0277\n",
      "[0/200][78/89]\tTotal Loss: 0.0218\tMAELoss: 0.0308\tMSSSIM Loss: 0.0200\n",
      "[0/200][79/89]\tTotal Loss: 0.0353\tMAELoss: 0.0316\tMSSSIM Loss: 0.0360\n",
      "[0/200][80/89]\tTotal Loss: 0.0210\tMAELoss: 0.0206\tMSSSIM Loss: 0.0211\n",
      "[0/200][81/89]\tTotal Loss: 0.0281\tMAELoss: 0.0348\tMSSSIM Loss: 0.0268\n",
      "[0/200][82/89]\tTotal Loss: 0.0262\tMAELoss: 0.0274\tMSSSIM Loss: 0.0259\n",
      "[0/200][83/89]\tTotal Loss: 0.0355\tMAELoss: 0.0318\tMSSSIM Loss: 0.0362\n",
      "[0/200][84/89]\tTotal Loss: 0.0317\tMAELoss: 0.0250\tMSSSIM Loss: 0.0330\n",
      "[0/200][85/89]\tTotal Loss: 0.0248\tMAELoss: 0.0314\tMSSSIM Loss: 0.0235\n",
      "[0/200][86/89]\tTotal Loss: 0.0372\tMAELoss: 0.0406\tMSSSIM Loss: 0.0365\n",
      "[0/200][87/89]\tTotal Loss: 0.0416\tMAELoss: 0.0506\tMSSSIM Loss: 0.0399\n",
      "[0/200][88/89]\tTotal Loss: 0.0337\tMAELoss: 0.0274\tMSSSIM Loss: 0.0349\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "0.001\n",
      "Epoch    22: reducing learning rate of group 0 to 5.0000e-04.\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "0.0005\n",
      "Epoch    34: reducing learning rate of group 0 to 2.5000e-04.\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "0.00025\n",
      "Epoch    61: reducing learning rate of group 0 to 1.2500e-04.\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "0.000125\n",
      "Epoch    72: reducing learning rate of group 0 to 6.2500e-05.\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "6.25e-05\n",
      "Epoch    93: reducing learning rate of group 0 to 3.1250e-05.\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "3.125e-05\n",
      "Epoch   111: reducing learning rate of group 0 to 1.5625e-05.\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "1.5625e-05\n",
      "Epoch   122: reducing learning rate of group 0 to 1.0000e-05.\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "1e-05\n",
      "Saved Intermediate: runs\\Rajaraman_ResNet\\177-20-20\\10KFold_217perEpoch\\network_intermediate_10.tar\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Training\n",
    "\n",
    "# Data augmentation transformations\n",
    "data_transforms = {\n",
    "    'train_transforms': tvtransforms.Compose([\n",
    "    custom_transforms.RandomAutocontrast(sample_keys_images, cutoff_limits=(0.1,0.1)),\n",
    "    custom_transforms.CenterCrop(sample_keys_images,256),\n",
    "    custom_transforms.RandomHorizontalFlip(sample_keys_images, 0.5),\n",
    "    custom_transforms.RandomAffine(sample_keys_images, degrees=10,translate=(0.1,0.1),scale=(0.9,1.1)),\n",
    "    custom_transforms.ToTensor(sample_keys_images),\n",
    "    custom_transforms.ImageComplement(sample_keys_images),\n",
    "    ])\n",
    "      ,\n",
    "    'test_transforms': tvtransforms.Compose([\n",
    "      custom_transforms.ToTensor(sample_keys_images),\n",
    "    ]),\n",
    "    }\n",
    "\n",
    "# K-FOLD VALIDATION\n",
    "splits=KFold(n_splits=split_k_folds,shuffle=True,random_state=42)\n",
    "foldperf={}\n",
    "\n",
    "print(target_key)\n",
    "dataset = datasets.JSRT_CXR(data_normal, data_BSE, transform=None)\n",
    "for fold, (train_idx,val_idx) in enumerate(splits.split(np.arange(len(dataset)))):\n",
    "    print('Fold {}'.format(fold + 1))\n",
    "    history = {'loss':[], 'ssim_acc':[]}\n",
    "    \n",
    "    # Subset sample from dataset\n",
    "    train_sampler = SubsetRandomSampler(train_idx)\n",
    "    test_sampler = SubsetRandomSampler(val_idx)\n",
    "    dl_training = DataLoader(datasets.JSRT_CXR(data_normal, data_BSE, transform=data_transforms['train_transforms']),\n",
    "                             batch_size=_batch_size, sampler=train_sampler, num_workers=0)\n",
    "    dl_validation = DataLoader(datasets.JSRT_CXR(data_normal, data_BSE, transform=data_transforms['train_transforms']),\n",
    "                             batch_size=_batch_size, sampler=test_sampler, num_workers=0)\n",
    "    \n",
    "    # Print example of transformed image\n",
    "    for count, sample in enumerate(dl_training):\n",
    "        image = sample[\"source\"][0,:]\n",
    "        image = torch.squeeze(image)\n",
    "        plt.imshow(image)\n",
    "        if count == 0:\n",
    "            break\n",
    "\n",
    "    ## Implementation of network and losses\n",
    "    input_array_size = (_batch_size, 1, image_spatial_size[0], image_spatial_size[1])\n",
    "    net = RajaramanModel.ResNet_BS(input_array_size)\n",
    "    # Initialise weights\n",
    "    net.apply(weights_init)\n",
    "\n",
    "    # Multi-GPU\n",
    "    if (device.type == 'cuda') and (ngpu > 1):\n",
    "        print(\"Neural Net on GPU\")\n",
    "        net = nn.DataParallel(net, list(range(ngpu)))\n",
    "    net = net.to(device)\n",
    "\n",
    "    # Optimiser\n",
    "    optimizer = optim.Adam(net.parameters(), lr=lr_ini, betas=(beta1, beta2))\n",
    "\n",
    "    # Learning Rate Scheduler\n",
    "    scheduler = optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode='min',\n",
    "                                                     factor=0.5, patience=int(10*epoch_factor), \n",
    "                                                     threshold=0.0001, min_lr=0.00001, verbose=True)\n",
    "    \n",
    "    # For each epoch\n",
    "    epochs_list = []\n",
    "    img_list = []\n",
    "    training_loss_list = []\n",
    "    reals_shown = []\n",
    "    #validation_loss_per_epoch_list = []\n",
    "    #training_loss_per_epoch_list = []\n",
    "    loss_per_epoch={\"training\":[], \"validation\":[]}\n",
    "    ssim_average={\"training\":[], \"validation\":[]}\n",
    "\n",
    "\n",
    "    # optionally resume from a checkpoint\n",
    "    if flag_load_previous_save:\n",
    "        if os.path.isfile(PATH_SAVE_NETWORK_INTERMEDIATE):\n",
    "            print(\"=> loading checkpoint '{}'\".format(PATH_SAVE_NETWORK_INTERMEDIATE))\n",
    "            checkpoint = torch.load(PATH_SAVE_NETWORK_INTERMEDIATE)\n",
    "            start_epoch = checkpoint['epoch_next']\n",
    "            reals_shown_now = checkpoint['reals_shown']\n",
    "            scheduler.load_state_dict(checkpoint['scheduler_state_dict'])\n",
    "            net.load_state_dict(checkpoint['model_state_dict'])\n",
    "            optimizer.load_state_dict(checkpoint['optimizer_state_dict'])\n",
    "            print(\"=> loaded checkpoint '{}' (epoch {}, reals shown {})\".format(PATH_SAVE_NETWORK_INTERMEDIATE, \n",
    "                                                                                start_epoch, reals_shown_now))\n",
    "            print(scheduler)\n",
    "        else:\n",
    "            print(\"=> NO CHECKPOINT FOUND AT '{}'\" .format(PATH_SAVE_NETWORK_INTERMEDIATE))\n",
    "            raise RuntimeError(\"No checkpoint found at specified path.\")\n",
    "    else:\n",
    "        print(\"FLAG: NO CHECKPOINT LOADED.\")\n",
    "        reals_shown_now = 0\n",
    "        start_epoch=0\n",
    "\n",
    "    # Loop variables\n",
    "    flag_break = False # when debugging, this will automatically go to True\n",
    "    iters = 0\n",
    "    net.train()\n",
    "    for param in net.parameters():\n",
    "        param.requires_grad = True\n",
    "    for epoch in range(start_epoch, total_num_epochs):\n",
    "        print(optimizer.param_groups[0]['lr'])\n",
    "        sum_loss_in_epoch = 0\n",
    "        for i, data in enumerate(dl_training):\n",
    "            # Training\n",
    "            net.zero_grad()\n",
    "            noisy_data = data[original_key].to(device)\n",
    "            cleaned_data = net(noisy_data)\n",
    "            loss, maeloss, msssim_loss = criterion_Rajaraman(cleaned_data, data[target_key].to(device))\n",
    "            loss.backward() # calculate gradients\n",
    "            optimizer.step() # optimiser step along gradients\n",
    "\n",
    "            # Output training stats\n",
    "            if epoch % 1000 == 0:\n",
    "                print('[%d/%d][%d/%d]\\tTotal Loss: %.4f\\tMAELoss: %.4f\\tMSSSIM Loss: %.4f'\n",
    "                      % (epoch, total_num_epochs, i, len(dl_training),\n",
    "                         loss.item(), maeloss.item(), msssim_loss.item()))\n",
    "            # Record generator output\n",
    "            #if reals_shown_now%(100*_batch_size)==0:\n",
    "            #    with torch.no_grad():\n",
    "            #        val_cleaned = net(fixed_val_sample[\"source\"].to(device)).detach().cpu()\n",
    "            #    print(\"Printing to img_list\")\n",
    "            #    img_list.append(vutils.make_grid(val_cleaned, padding=2, normalize=True))\n",
    "            #iters +=1\n",
    "            #if flag_debug and iters>=10:\n",
    "            #    flag_break = True\n",
    "            #    break\n",
    "\n",
    "            # Running counter of reals shown\n",
    "            reals_shown_now += _batch_size\n",
    "            reals_shown.append(reals_shown_now)\n",
    "\n",
    "            # Training loss list for loss per minibatch\n",
    "            training_loss_list.append(loss.item()) # training loss\n",
    "            sum_loss_in_epoch += loss.item()*len(cleaned_data)\n",
    "        # Turn the sum loss in epoch into a loss-per-epoch\n",
    "        loss_per_epoch[\"training\"].append(sum_loss_in_epoch/len(train_idx))\n",
    "\n",
    "\n",
    "        with torch.no_grad():\n",
    "            # Training Accuracy per EPOCH\n",
    "            ssim_training_list = []\n",
    "            for train_count, data in enumerate(dl_training):\n",
    "                noisy_training_data = data[original_key].to(device)\n",
    "                true_training_data = data[target_key]\n",
    "                cleaned_training_data = net(noisy_training_data)\n",
    "\n",
    "                for ii, image in enumerate(cleaned_training_data):\n",
    "                    clean_training_numpy = image.cpu().detach().numpy()\n",
    "                    true_training_numpy = true_training_data[ii].numpy()\n",
    "                    clean_training_numpy = np.moveaxis(clean_training_numpy, 0, -1)\n",
    "                    true_training_numpy = np.moveaxis(true_training_numpy, 0, -1)\n",
    "                    ssim_training = skimage.metrics.structural_similarity(clean_training_numpy, true_training_numpy, multichannel=True)\n",
    "                    ssim_training_list.append(ssim_training) # SSIM per image\n",
    "            ssim_average[\"training\"].append(np.mean(ssim_training_list))\n",
    "\n",
    "            # Validation Loss and Accuracy per EPOCH\n",
    "            sum_loss_in_epoch =0\n",
    "            ssim_val_list = []\n",
    "            for val_count, sample in enumerate(dl_validation):\n",
    "                noisy_val_data = sample[original_key].to(device)\n",
    "                cleaned_val_data = net(noisy_val_data)\n",
    "\n",
    "                # Loss\n",
    "                true_val_data = sample[target_key]\n",
    "                val_loss, maeloss, msssim_loss = criterion_Rajaraman(cleaned_val_data, true_val_data.to(device))\n",
    "                sum_loss_in_epoch += val_loss.item()*len(cleaned_val_data)\n",
    "\n",
    "                # Accuracy\n",
    "                for ii, image in enumerate(cleaned_val_data):\n",
    "                    clean_val_numpy = image.cpu().detach().numpy()\n",
    "                    true_val_numpy = true_val_data[ii].numpy()\n",
    "                    clean_val_numpy = np.moveaxis(clean_val_numpy, 0, -1)\n",
    "                    true_val_numpy = np.moveaxis(true_val_numpy, 0, -1)\n",
    "                    ssim_val = skimage.metrics.structural_similarity(clean_val_numpy, true_val_numpy, multichannel=True)\n",
    "                    ssim_val_list.append(ssim_val) # SSIM per image\n",
    "            # After considering all validation images\n",
    "            loss_per_epoch[\"validation\"].append(sum_loss_in_epoch/len(val_idx))\n",
    "            ssim_average[\"validation\"].append(np.mean(ssim_val_list))\n",
    "        epochs_list.append(epoch)\n",
    "        # LR Scheduler after epoch\n",
    "        scheduler.step(val_loss)\n",
    "\n",
    "        # Save the network in indications\n",
    "        #if epoch % 5 == 0:\n",
    "        #    if not flag_debug:\n",
    "        #        #torch.save(net.state_dict(), PATH_SAVE_NETWORK_INTERMEDIATE)\n",
    "        #        torch.save({\n",
    "        #        'epochs_completed': epoch+1,\n",
    "        #        'epoch_next': epoch+1,\n",
    "        #        'model_state_dict': net.state_dict(),\n",
    "        #        'optimizer_state_dict': optimizer.state_dict(),\n",
    "        #        'loss': loss,\n",
    "        #        'scheduler_state_dict': scheduler.state_dict(),\n",
    "        #        'reals_shown': reals_shown_now\n",
    "        #        }, PATH_SAVE_NETWORK_INTERMEDIATE)\n",
    "        #        print(\"Saved Intermediate: \"+ str(PATH_SAVE_NETWORK_INTERMEDIATE))\n",
    "        #if flag_break:\n",
    "        #    break\n",
    "        \n",
    "    #After all epochs, save results:\n",
    "    history[\"loss\"] = loss_per_epoch\n",
    "    history[\"ssim_acc\"] = ssim_average\n",
    "    \n",
    "    foldperf['fold{}'.format(fold+1)] = history\n",
    "    \n",
    "    # Final Save for each fold\n",
    "    if not flag_debug:\n",
    "        torch.save({\n",
    "                'epochs_completed': epoch+1,\n",
    "                'epoch_next': epoch+1,\n",
    "                'model_state_dict': net.state_dict(),\n",
    "                'optimizer_state_dict': optimizer.state_dict(),\n",
    "                'loss': loss,\n",
    "                'scheduler_state_dict': scheduler.state_dict(),\n",
    "                'reals_shown': reals_shown_now\n",
    "                }, PATH_SAVE_NETWORK_INTERMEDIATE.format(fold+1))\n",
    "        print(\"Saved Intermediate: \"+ str(PATH_SAVE_NETWORK_INTERMEDIATE.format(fold+1)))\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2a437c08",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training complete\n",
      "Performance of 10 fold cross validation\n",
      "Average Training Loss: 0.018 \t Average Test Loss: 0.017 \t Average Training Acc: 0.976 \t Average Test Acc: 0.976\n",
      "Best fold: 5\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"Training complete\")\n",
    "# Find the best fold\n",
    "\n",
    "\n",
    "# AVERAGE PERFORMANCE\n",
    "testl_f,tl_f,testa_f,ta_f=[],[],[],[]\n",
    "k=10\n",
    "for f in range(1,k+1):\n",
    "    tl_f.append(np.mean(foldperf['fold{}'.format(f)]['loss']['training']))\n",
    "    testl_f.append(np.mean(foldperf['fold{}'.format(f)]['loss']['validation']))\n",
    "\n",
    "    ta_f.append(np.mean(foldperf['fold{}'.format(f)]['ssim_acc']['training']))\n",
    "    testa_f.append(np.mean(foldperf['fold{}'.format(f)]['ssim_acc']['validation']))\n",
    "\n",
    "print('Performance of {} fold cross validation'.format(k))\n",
    "print(\"Average Training Loss: {:.3f} \\t Average Test Loss: {:.3f} \\t Average Training Acc: {:.3f} \\t Average Test Acc: {:.3f}\"\n",
    "      .format(np.mean(tl_f),np.mean(testl_f),np.mean(ta_f),np.mean(testa_f)))\n",
    "print(\"Best fold: {}\".format(np.argmax(testa_f)+1))\n",
    "\n",
    "# Averaging accuracy and loss\n",
    "diz_ep = {'train_loss_ep':[],'test_loss_ep':[],'train_acc_ep':[],'test_acc_ep':[]}\n",
    "for i in range(total_num_epochs):\n",
    "    diz_ep['train_loss_ep'].append(np.mean([foldperf['fold{}'.format(f+1)]['loss']['training'][i] for f in range(k)]))\n",
    "    diz_ep['test_loss_ep'].append(np.mean([foldperf['fold{}'.format(f+1)]['loss']['validation'][i] for f in range(k)]))\n",
    "    diz_ep['train_acc_ep'].append(np.mean([foldperf['fold{}'.format(f+1)]['ssim_acc']['training'][i] for f in range(k)]))\n",
    "    diz_ep['test_acc_ep'].append(np.mean([foldperf['fold{}'.format(f+1)]['ssim_acc']['validation'][i] for f in range(k)]))\n",
    "\n",
    "# Plot losses\n",
    "plt.figure(figsize=(10,8))\n",
    "plt.semilogy(diz_ep['train_loss_ep'], label='Train')\n",
    "plt.semilogy(diz_ep['test_loss_ep'], label='Validation')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "#plt.grid()\n",
    "plt.legend()\n",
    "if not flag_debug:\n",
    "    plt.savefig(os.path.join(output_save_directory, current_date + \"_loss\"+\".png\"))\n",
    "plt.show()\n",
    "\n",
    "# Plot accuracies\n",
    "plt.figure(figsize=(10,8))\n",
    "plt.semilogy(diz_ep['train_acc_ep'], label='Train')\n",
    "plt.semilogy(diz_ep['test_acc_ep'], label='Validation')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('SSIM')\n",
    "#plt.grid()\n",
    "plt.legend()\n",
    "if not flag_debug:\n",
    "    plt.savefig(os.path.join(output_save_directory, current_date + \"_accuracy\"+\".png\"))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "3e0f3956",
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'fixed_val_sample' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-6-b0e89ac468dd>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     38\u001b[0m \u001b[1;31m# Final Model:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     39\u001b[0m \u001b[1;32mwith\u001b[0m \u001b[0mtorch\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mno_grad\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 40\u001b[1;33m     \u001b[0minput_image\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfixed_val_sample\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'source'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     41\u001b[0m     \u001b[0minput_images\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mvutils\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmake_grid\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0minput_image\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpadding\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnormalize\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mTrue\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     42\u001b[0m     \u001b[0mtarget_images\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mvutils\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmake_grid\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfixed_val_sample\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'boneless'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpadding\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnormalize\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mTrue\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'fixed_val_sample' is not defined"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAFOCAYAAADdMIrBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3jb5NoG8PtJ0qQr3enem+7N6KCLTvbeh3WAw4Gyoeyyy/zYlD0KHCijFGihG2ihm+690pXulbZp2oz3+8OyI8uSLA/ZTnL/rstXbFm2Xiuy/OgdzytKKRARERFRYkmKdwGIiIiIKBCDNCIiIqIExCCNiIiIKAExSCMiIiJKQAzSiIiIiBIQgzQiIiKiBMQgjYhKBBG5TkRmR/D6q0RkSjTLFE0iMkZEHov2ukSUuIR50ogoEiKSCeAmpdS0OJfjOq0cveJZDjOJso+IqHhhTRoRlXoiklKat09EiYlBGhG5QkTSROQ1EcnSbq+JSJr2XA0R+UVEDonIARGZJSJJ2nMPisgOETkiImtFZIDF+1cXkZ9EJFtE5gNopnuusYgoffAjIr+LyE3a/etE5C8R+T8ROQBglLG5VHv9rSKyXkQOisjbIiLac8ki8oqI7BORzSJyu3F7uvcZC6AhgJ9F5KiIPKAr340ishXADG3db0Vkl4gcFpE/RaSt7n0+FZFntPt9RWS7iNwrIntEZKeIXB/mutVF5GdtPy4QkWciaTYmouhhkEZEbnkEwGkAOgHoCKAHgEe15+4FsB1ABoBaAB4GoESkFYDbAXRXSqUDGAwg0+L93waQC6AOgBu0WyhOBbAJQE0Az1qsczaA7lr5L9XKAwD/BjBU+2xdAJxvtRGl1DUAtgI4RylVUSn1ou7pMwGconvfXwG00Mr0D4AvbcpfG0BlAPUA3AjgbRGpGsa6bwM4pq3zL+1GRAmAQRoRueUqAE8ppfYopfYCeBLANdpzefAEV42UUnlKqVnK00G2AEAagDYiUkYplamU2mh8YxFJBnARgMeVUseUUisAfBZi+bKUUm8qpfKVUsct1hmtlDqklNoKYCY8QRngCdheV0ptV0odBDA6xG17jdLKfxwAlFIfK6WOKKVOABgFoKOIVLZ4bR48+zdPKTUJwFEArUJZV7cfn1BK5SilViH0/UhELmGQRkRuqQtgi+7xFm0ZALwEYAOAKSKySURGAoBSagOAu+AJUPaIyNciUheBMgCkANhmeP9QbAu+Cnbp7ucAqKjdr2t4vZP3si2D1oQ6WkQ2ikg2imoQa1i8dr9SKt+ifE7XNduP4X4WIooyBmlE5JYsAI10jxtqy6DVFt2rlGoK4BwA93j7nimlvtJGaDYCoAC8YPLeewHkA2hgeH+vY9rf8rpltQ3vEcnQ9p0A6useN7BaMci29MuvBHAegIHwNE021pZLGOVzyrsfQ/ksRBQjDNKIKBrKiEhZ3S0FwP8APCoiGSJSA8DjAL4AABE5W0Saax3xs+Fp5iwQkVYi0l8bYJAL4Lj2nB+lVAGAH+Dp8F9eRNpA15dKa17dAeBqrYbqBugGFkTBOAB3ikg9EakC4MEg6+8G0DTIOukATgDYD09w+VzEpQzCZD+2BnCt29slImcYpBFRNEyCJ6Dy3kYBeAbAQgDLACyHpyP8M9r6LQBMg6dv1BwA7yilfoenP9poAPvgaWqsCc+gAjO3w9NktwvApwA+MTz/bwD3wxP0tAXwdyQf0OADAFPg+WyL4fn8+TAJKDXPwxOwHhKR+yzW+RyeJtsdAFYBmBvF8tq5HZ6au10AxsITXJ+I0baJyAaT2RIRRUhEhgIYo5RqFHTlBCciLwCorZTiKE+iOGNNGhFRiESknIgME5EUEakH4AkA4+NdrnCISGsR6SAePeBJ0VEsPwtRScMgjYgodAJPSpGD8DR3roanz11xlA5Pv7Rj8PS1ewXAhLiWiIgAsLmTiIiIKCGxJo2IiIgoATFIIyIiIkpAAZMBlwQ1atRQjRs3jncxiIiIiIJatGjRPqVUhnF5iQzSGjdujIULF8a7GERERERBiYjptHZs7iQiIiJKQAzSiIiIiBIQgzQiIiKiBMQgjYiIiCgBMUgjIiIiSkAM0oiIiIgSEIM0IiIiogTEII2IiIgoATFIIyIiIkpADNLCMHfTfizZdijexSAiIqISjEFaGJ78eRXemrEh3sUgIiKiEoxBWhgEAKDiXAoiIiIqyRikhUEEUIzRiIiIyEUM0sIgwno0IiIicheDtDAIBIpVaUREROQiBmlhYE0aERERuY1BWhgk3gUgIiKiEo9BWpjY2klERERuYpAWDhE2dxIREZGrGKSFQQAOHCAiIiJXMUgLg7BTGhEREbmMQVoYPDVp8S4FERERlWQM0sIgIlDslUZEREQuYpAWBtakERERkdsYpIWBc3cSERGR2xikhUHA5k4iIiJyF4O0cLAmjYiIiFzGIC0MAs7dSURERO5ikBYGYZRGRERELmOQFgb2SSMiIiK3MUgLA0d3EhERkdsYpIVBhK2dRERE5C4GaWEQCCdYJyIiIlcxSAsDa9KIiIjIbQzSwsSKNCIiInITg7QweCZYJyIiInIPg7QwCMCqNCIiInIVg7QwsE8aERERuY1BWhgErEgjIiIidzFIC4OIxLsIREREVMIxSAsTp4UiIiIiNzFICwObO4mIiMhtDNLCwLk7iYiIyG0M0sLCPGlERETkLgZpYfDUpDFMIyIiIvcwSAsDx3YSERGR2xikhYF90oiIiMhtDNLCIBCm4CAiIiJXMUgLA2vSiIiIyG0M0sLAuTuJiIjIbQzSwiAQju4kIiIiV6XEY6MiUg3ANwAaA8gEcKlS6qDJepkAjgAoAJCvlOoWu1LaYE0aERERuSxeNWkjAUxXSrUAMF17bKWfUqpTwgRo0FJwMEojIiIiF8UrSDsPwGfa/c8AnB+ncoRFhDMOEBERkbviFaTVUkrtBADtb02L9RSAKSKySERujlnpgvBMsM4wjYiIiNzjWp80EZkGoLbJU4+E8DY9lVJZIlITwFQRWaOU+tNiezcDuBkAGjZsGHJ5Q8HRnUREROQ214I0pdRAq+dEZLeI1FFK7RSROgD2WLxHlvZ3j4iMB9ADgGmQppR6H8D7ANCtWzdXYyhPTZqbWyAiIqLSLl7NnT8B+Jd2/18AJhhXEJEKIpLuvQ9gEIAVMSuhDU+fNEZpRERE5J54BWmjAZwlIusBnKU9hojUFZFJ2jq1AMwWkaUA5gOYqJT6LS6lNWBNGhEREbktLnnSlFL7AQwwWZ4FYJh2fxOAjjEumjOcFoqIiIhcxhkHwlBYqLDj0PF4F4OIiIhKMAZpYfhxSRYAYE92bpxLQkRERCUVg7QIHDqeF+8iEBERUQnFII2IiIgoATFIIyIiIkpADNKIiIiIEhCDNCIiIqIExCCNiIiIKAExSCMiIiJKQAzSiIiIiBIQg7QISLwLQERERCUWgzQiIiKiBMQgLQKcY52IiIjcwiCNiIiIKAExSIsA+6QRERGRWxikERERESUgBmlERERECYhBGhEREVECYpBGRERElIAYpBERERElIAZpRERERAmIQRoRERFRAmKQFgFhojQiIiJyCYO0CCjOC0VEREQucRykiUgFEUl2szBERERE5GEZpIlIkohcKSITRWQPgDUAdorIShF5SURaxK6YRERERKWLXU3aTADNADwEoLZSqoFSqiaA3gDmAhgtIlfHoIwJi33SiIiIyC0pNs8NVErlGRcqpQ4A+B7A9yJSxrWSEREREZViljVp3gBNRJqJSJp2v6+IjBCRKvp1SisOHCAiIiK3OBk48D2AAhFpDuAjAE0AfOVqqYoJNncSERGRW5wEaYVKqXwAFwB4TSl1N4A67haLiIiIqHRzEqTlicgVAP4F4BdtGfuiAVi89VC8i0BEREQllJMg7XoApwN4Vim1WUSaAPjC3WIVD4ePl+oueUREROQiu9GdAACl1CoAIwBARKoCSFdKjXa7YERERESlWdCaNBH5XUQqiUg1AEsBfCIir7pfNCIiIqLSy0lzZ2WlVDaACwF8opTqCmCgu8UiIiIiKt2cBGkpIlIHwKUoGjhAYJ40IiIico+TIO0pAJMBbFRKLRCRpgDWu1us4kGBURoRERG5w8nAgW8BfKt7vAnARW4WioiIiKi0czJwoL6IjBeRPSKyW0S+F5H6sShcomNzJxEREbnFSXPnJwB+AlAXQD0AP2vLSj3GaEREROQWJ0FahlLqE6VUvnb7FECGy+UqFpI5eScRERG5xEmQtk9ErhaRZO12NYD9bhesOCiXmhzvIhAREVEJ5SRIuwGe9Bu7AOwEcDE8U0URERERkUucjO7cCuBc/TIReRnAfW4VioiIiKi0c1KTZubSqJaimLmlT1MAQJu6leJcEiIiIiqpwg3SSnWP+c4NqwAAypVhnzQiIiJyh2VzpzahuulTKOVBmhfzpBEREZFb7PqkLYInFZhZQHYyko2KyCUARgE4BUAPpdRCi/WGAHgdQDKAD5VSoyPZbvR4dgmnhSIiIiK3WAZpSqkmLm53BYALAbxntYKIJAN4G8BZALYDWCAiPymlVrlYLkeYHo2IiIjcZtknTUQa271QPMKaHkoptVoptTbIaj0AbFBKbVJKnQTwNYDzwtmeW9jcSURERG6xa+58SUSSAEyAp+lzL4CyAJoD6AdgAIAn4KnlckM9ANt0j7cDONWlbYWEFWlERETkNrvmzktEpA2Aq+BJaFsHwHEAqwBMAvCsUirX6vUiMg1AbZOnHlFKTXBQNrNYyLLuSkRuBnAzADRs2NDB2xMRERElLttktlr/r0fCeWOl1MCwSlRkO4AGusf1AWTZbO99AO8DQLdu3VxtiBStUxqbO4mIiMgttkGaiFQGMASepkcFT5A0WSl1KAZlWwCghYg0AbADwOUArozBdoNicycRERG5zW7gwLUA/gHQF0B5ABXg6Yu2SHsubCJygYhsB3A6gIkiMllbXldEJgGAUiofwO0AJgNYDWCcUmplJNuNNqbgICIiIrfY1aQ9AqCrsdZMRKoCmAfg83A3qpQaD2C8yfIsAMN0jyfB0/8toTAFBxEREbnNbloogXlH/UKwxQ8A+6QRERGRe+xq0p4F8I+ITEFRKoyG8CSXfdrtgiUy1qQRERGR2yxr0pRSnwHoBuAPACfgmQrqdwDdlFKfxqJwRERERKVVsBQcB+HJ9E8m2NpJREREbrHrk2ZJRJZHuyDFibBLHhEREbnMsiZNRC60egrmMwmUOoojB4iIiMglds2d3wD4EuatemXdKU4xwYo0IiIicpldkLYMwMtKqRXGJ0Qk0imfSgTWoxEREZFb7Pqk3QUg2+K5C1woS7HBijQiIiJym2VNmlJqls1zC90pDhEREREBQVJwAICIvGGy+DCAhUqpCdEvUvHBcQNERETkFicpOMoC6ARgvXbrAKAagBtF5DUXy5awhFMOEBERkcuC1qQBaA6gv1IqHwBE5F0AU+CZHqpU50vj0AEiIiJyi5OatHoAKugeVwBQVylVAM90UaUO69GIiIjIbU5q0l4EsEREfocnPukD4DkRqQBgmotlS3jsk0ZERERuCVqTppT6CMAZAH7Ubr2UUh8qpY4ppe53u4CJqECLzu76ZkmcS0JEREQllZOaNADoDqC3dr8AQJY7xSkeVu/0pI/bfvB4nEtCREREJVXQmjQRGQ3gTgCrtNsIEXne7YIlMk6wTkRERG5zUpM2DEAnpVQhAIjIZwAWA3jIzYIRERERlWZORncCQBXd/cpuFKQ4YZo0IiIicpuTmrTnASwWkZkoGt3JWjQiIiIiFwUN0pRS/9PSb3SHJ0h7EEAjl8uV0Kwq0uZt2o99R09ieIc6MS0PERERlTyORncqpXYC+Mn7WETmA2joVqGKq8venwsAGN5heJxLQkRERMWd0z5pRqW6VxZz2BIREZHbwg3SSnWcUqojVCIiIooJy+ZOEfkZ5sGYAKjuWomKgVIdoRIREVFM2PVJeznM54iIiIgoQpZBmlLqj1gWhIiIiIiKhNsnrVTbk33Cd/+PdXvjWBIiIiIqqRikhWFB5gHf/axDnGSdiIiIoo9BWhhO5hfGuwhUDO07egKNR05k7SsRETkSNJmtxSjPwwAWAnhPKZXrRsESWbfGVbF29xEATMdBzi3fcRgA8PHszTizZUacS0NERInOSU3aJgBHAXyg3bIB7AbQUntc6gxqWzveRSCd3dm5yCtg7SYREZUsTqaF6qyU6qN7/LOI/KmU6iMiK90qWCJrU6eS7fO5eQVITU5CUhLr2dyWm1eAU5+bjgu71MOrl3aKd3GIiIiixklNWoaI+Obp1O7X0B6edKVUxcjIH5aj70szkZtX4FvW+rHf8Nyk1XEsVelxIs9TgzZt1e44l4SI3KaUwsuT12IHB2xRKeEkSLsXwGwRmSkivwOYBeB+EakA4DM3C1dcZO7PwfaD/ieNz+dsiVNpKGFxqgqiiKzeeQRvzdyA2778J95FIYqJoM2dSqlJItICQGt4+smv0Q0WeM3NwhVnJ9lHiiwIW8GJwlKoPFc6eRxhT6WEkz5pANAVQGNt/Q4iAqXU566Vqhh6bdq6eBeBiIiISpCgzZ0iMhaeuTp7Aeiu3bq5XK5iZ/LKXWG/9rcVu5CdmxfF0pQcs9bvRT5rJYmIqBRy0ietG4CeSqnblFJ3aLcRbhesuBGTjGk7Dx/HZe/NwaqsbIyduwVvTl8fsE7mvmO49YtFuHfc0lgUs1iZvX4frvloPt6euTHourHu7rVu9xFk7jsW460GGv7GLHR9emq8i0FERC5w0ty5AkBtADtdLkuxkVbGJLY16Wd0+vMzAADD3pjlW3bHgBZ+6+Sc9IwK3XYgJ3oFLCH2HPF0fczcH/9gyGjQ//0JAMgcPTyu5ViZlR2wrKDQE7ImMwUMEdno+vRUXNGjIe4b3CreRSELTmrSagBYJSKTReQn783tgiWySmXLBCwL5ecwr6AQSnGoXzQxHCnSYdRk9H5hRryLQSXczsNMg1Hc7T92Em/N3BDvYvjMXLsHt4xdGO9iJBQnNWmj3C5ESXDC4Wijx35cgbFzPek53r6yC5rUqOBmsSxtP5iDtJRkZKSnxWX7gKfGRwAm/Y2yYycLcOxkQfAVydLYOZno3LAq2tWrHO+iJKQZa3bjhk8X4sNru2Fgm1rxLg6VENd/siDeRUg4QWvSlFJ/mN1iUbiSyBugAcB/v4pNrp8dh45jnTbXqFevF2ai+7PTYrJ9K80enuTXFGzEykaKl8cmrMTZb86OdzES1rLtnnlol2nz0RKROyyDNBGZrf09IiLZutsREQnsCEMJq+foGb4+VIlmza4jQddxUs9W0uK5xyeswO0uBPGP/rgcjUdOjPr7ElH4Fm05iN9WhJ8hgEouyyBNKdVL+5uulKqku6UrpewnryTHdmd7Osfrg5W/N+7DWa/+gXELt2H66t0BtWDh+uzvTBQWlrBwphi1lCqbUHLzvmN+/RQ/n7MFvyyL/lidL+Zujfp7UunDWu7ouujdv3HrF4viXQxKQE4GDkBEkkWkrog09N7cLlhpcf2ngW3wT0xYifV7juKB75bhxs8W2taCKaUwffVuR7nEnvhpJX5eloXDOcUrJ5vt70Ex/LEwxpUrsw6j38u/4/0/N8WlPEThcuMaSSmF8Yu34yRnFSBylMz2DgC7AUwFMFG7/eJyucih39ftxY2fLcSbMzwjdDbvO4at+63TeRw7UYCOT01xtUx7j5zAi7+tiWmtXTGqUAuw7YBnlNyiLQfjXBKi+Juyajfu/mYpZ3EhgrOatDsBtFJKtVVKtdduHdwuWGk0YckOy+f+/flCDH9jFo7oZib4aWkWZqzeAwB4ffp6rNt9BP1e/h19XpoJAOj45BSc+1Z0Oj+/OmWt475MI79fhnd+34g5m/ZHZdvFOQCz898v/0HvF5kqI9ZW7Dhc8pr9S5DDxz3nuD1HTkT9vbfsP4bnf11dolMgbdx7tER/vtLGSZC2DUBUh/CIyCUislJECkXEcoopEckUkeUiskRESnzylDu/XoLd2bmmLXhTV+3GyqxszNlYFPiM+N9iv9GiT/28yu81h4/n+UZheYU7ufcbM5zn0vGmIyl0+URxzUfz8NmcTFe34aaJy3f6atGA2Lfc/vDP9rgmUS4sVFiYeSCm25y3aT/OfnM2Pv5rc0y3G64T+QVoPHIiXp2yNuL3OnYiHyuzQjuVb9p7FLl5MU7n4uIX4Zaxi/DeH5uwce9R9zYSRwszD2DAK3/gC93vAhVvToK0TQB+F5GHROQe7y3C7a4AcCEAJ0MO+ymlOimlSsV8oa9NW4ftB61/OBWAcQu3YUWQoe8FCVBTcDQ33+/xih2H0XjkRGw7kBOVRJiz1u/Dq1M9TSLhftpXpqzFmVrNo96hnJM4eOxkBKVzxy/LstB45MSIy6aUwj3jluLCd/+OUslC98GsTbh4zBzMXr8vZtvcdtBz3K3aWTwGqOec8ARIn0fhR9dTGz8beQ7nws3NK0D/V/7AnV8vjnjb4YikBv2R8cvRftTkgOVOP3txtUmbqm7p9uKdGsWsJrC09lF0EqRthac/WiqAdN0tbEqp1UqpyC8NS6D/zd+G3Dzrg/Hw8Tw88N2yoDmc7hm3JKTtKqXQ/dlp+GaBs9F/uw7n4j9fLELOyfyA57IOeX4I7zaU4WvtvX9fu8c3ZZb3sWmZHJUkMm/O2IAtJn34Oj01FZ0TcE7MT/7KBICwagL2aCOJ9fa60KTk1Po9ns+QFY/M9fG/honI/+ZvxWbd3LFrdmUHrfFamOnp8+i0hvukFtD8vSGw20Ki7b6eo2fgBt0grC/nbcWR3MBzk1dxbQ38bcUuNB45Edm5iTX4a82ubNPzi1NKKcsm2gWZB/DatHVo+eivWB+lTAfFSdAZB5RST8aiIFabBzBFRBSA95RS78exLAnhge+WOVpvwpKskN63UHl+sB/6YTku6x588O4Lv63Bryt24aw2tXBhl/p+zx3XfiyMwWaS1tZqrOS77pMFcZ8D06nd2bmWQWW4wm2CDlWP56bHZkMJzLuri8tvtFU5H/phOdLTUrD8ycE4eOwkhrw2C+d2rIs3rugc0/K5cezapaqxsuPQcew4FDzYl1h92Vzyzu/aALG9x9CxQZXAFSx23Z/r9mL66t148rx2rpRryGuzkJqchHXPDg3r9c0enoRWtc0ze10yZo7v/sqsbLSoFXod0V8b9iG/UOHMlhlhlS+e7JLZvqb9/Vk/Z6fTuTtFZJqIrDC5nRdC+XoqpboAGArgvyLSx2Z7N4vIQhFZuHfv3hA2UbpFoybl8Qkr0HjkRIydkxnwnL5mrihIc3YSToTT6W8rdqLxyIk4esJzVX7Dpwvw4PfLY7b9bxduM12eCEHGrPV7Q24+2nU41/djWlxrM6Llz3V7cSjHWbO1/rvgzZt4RDsmj2m12cVxdPDGvUeLZbmLm2s/no/P5kTeZL5+9xE8O3GVeXOk4VyQm1eAuQ4HjxUqYLWDLgjhBPAAcNWH8/Cvj+eH9dp4s2vuHKv9fRnAKyY3W0qpgUqpdia3CU4Lp5TK0v7uATAeQA+bdd9XSnVTSnXLyCh+0XKsGAOf7s9Ow/6jzgO1n5b619DN2bgfn2tf/scmrAzYhj6g8V7EutFdzvi5CgsVxi3YFnEflNemrff8nboOu7NzsS+EfWXkPa/NXLvXcnRh1qHjfs/d77DmNNYWbTmIaz6aj5cmh9Zr4bTnp6PnaP8RrYkQjLtt+Buz8MSEFb7H2bl5uPbj+fj356GNh/p97Z6IZw+Zump3wI/s2zM3hFZDHKUIe8Arf+CiMPpFOq0Qe2nyGr/97lXKrw/Cdu3H8/HBrM3Y5aBp88mfV+Ly9+diw57S10QZTXYzDizS/sZl7k4RqSAi6d77AAbBM+CALDi5ynhm4uqAZdNXF52cCxUwbkFR7c2xE/k4kV/U12XE/zydiL0n+a0mowOPW/SN8dak2Q0PX5B5wLS/Rdenp+Jlm4DA+I4/LN6BB75fhp6jZ+CreUW1eZNX7sKy7Ycs38fKh7M3418fz7ft5xKKHxYHplvZvO8Yzhg9A+/+sdFveX5BIb6ct8VvMEh+QWQ/MzPWRNZk6w3sN+09FmTNxOL9cY91ioKVWdl+NRl5WifojSHuvw17rPsiOv1Mt3+1GL+v9W9teGnyWlwXwuTWOw+H3//ITLdnpvmdZ4DoNKW+PXOj336PxQVBzsl830wy8eJr1lfKcW2WE95WEHGwJ9dqs+h4U6qEwupQLo21706S2bYQke9EZJWIbPLeItmoiFwgItsBnA5goohM1pbXFZFJ2mq1AMwWkaUA5gOYqJT6LZLtlnR/mXTwNfI22+m9OHmN3+MHvi+qvWn7xGRc+I7/le62AznwxgjzNgemUDhoMaNBkq8mLfCbVlCocPxkAS4ZMwcdRk3xdCTVPb//2Em8NXMDnv5llaPZFbzNSHuOnMDD4z21ed8s2Ipbxi7CuW/9BQA44/nQ+mit2XUEOScDA9Av5m4JOUXCUZNAdIc28lCfZgUAPv5rMx4ZvwJfzd/qC9RG/hBZDZsxNUtJd9pz03HvuKXxLoYrnPaz0n+j9jsYHWz1g/jr8p34dtF2R9v02ro/B4+MX2456nzf0RPYk+1fS519PDoXRE79tmIndkUh+Lz8/bk4NUH6f46duwWXvz833sWgCDgZ3fkJgHcB5APoB+BzFDWFhkUpNV4pVV8plaaUqqWUGqwtz1JKDdPub1JKddRubZVSz0ayTbJ2KCfP9ip8ZZZ/X4HeL87Ez0udD0zwXkmJxcABAHj6l1XIKywKvvxGL+p+gz6avRlTV+0Ouk3jx1m67VBAX7Is3Ql5T3Yuxln0/wrm0R9XhJRHzopVTag36M0+nuf7kTMbkRratqJn9vp9+Hp+6HOChtu/JBy7snPx/T+hBRYlTZ6h9rX3izPwf1Ots/rra6ABTw6u3LwCLN7mXxM9+tc1aPO4//VzQaHy62pwx9eL8eW8rVjqoBbbm2rht5Wxm3C8sFDh1i/+waXveTqpz9u0P+z8cPG8ADJ+p9yq6Q7luxvN2i/ve6UUaUIAACAASURBVM3btB8/lJLvs5MgrZxSajoAUUptUUqNAtDf3WJRLOUXKhQYvklrdoWXR8qslsubQNN7wb/ApPbt1xU7/SrQH/5hBX5d7plg3DjXaLA+bSfzC/HsJP9m3fPe/sv2Ndd9sgAPfLcsoM+ZfuL7YD79a3NYzWhOm2AO5ZwMOugiklxCi7YcdBQA+1O4+qN5GPlD4GCK/3yxCLeMNe9zNU23HX1N0NET+bjty0WupQYJVunkpH+mUgofz94cUcqBaDl8PM/Xx894ZOQVFNoej9sOHMfr09dbPv/Cb2t06+bg4jFz8Mh4/x4nAsGYPzYG1DBfMuZvtHjk14D3dHKsu5XjcduBHF/aF6N9xzz/9+0Hc7Bl/zFc9v5cXw18rF307t+4/hPzTu7xbu5z0szpW1f3Zbvv26W45qN5UUtefdn7c3FPCa0ZN3ISpOWKSBKA9SJyu4hcAKCmy+WiGDNeZQ95bVZYCWdH/7rG8jlvn7TpJv2hlPL/Us/PPOBb74BhBJzpVZxukZNRQkZ7tR/nUH8gfltRdLU/6udVWLv7CPYeOYErP5gbkMtsuUUC4mBb9O6VD2ZtxiGLpuRVWm2nsc/S/d8uxS/LTGo9dWf7E/kFGDt3Cy569++QO7Lb+XXFLkxeuRtj52QG9De6SbedrQdy0PXpqdh+MAc//LMdk5bvwhs2wYOeUsqvCf/tmRt8CY69zJr4zfb5jDW70fWZaZi13n50+KZ9x/DUL6tw25f/OCpjJIIF/Vbf0fyCQrR45Fc8/UtgH1QAIfXLPHIi31cb/v0/2x01Cf6z1fP+jUdOxHyTizIz3o+qPw+MW7gNSilc/eE8xxcQZts7eiIf43X9QI279Z5vPD/4haqomXWdCzm5vlmwFZNtagjnbtqPRVsOYubavbYXDPqLjcPH8wIGdAVzKOckfjTpF+vU2l1H8NTP5qM8zXy3aDtmrd+Hi3XpNOw4PQuv330Ew9+YFfW8cQWFCjd8usDxsesmJ0HaXQDKAxgBoCuAqwH8y81CUey9/2dgN0OrgMDOh7MDp9vxfo+9NWOhWrw1+A/KEd0PcTgXm+F2KL71i0V+j9+Yvh7dn52Gvzfux/OT1mBB5gG0HzUZh3PyfCNFgcDh6oCzwQBWJ6Nhb8wyXf7tou24/avAjPH6Lb01YwMe+9F+TM6xE/kWtXRFe86qeeixCSvxlllzsFaIcQu2Yf+x0H40sg4dx4ezNuGtGRvQ7onJvh+0lyavDQjwXp1SFLTpawLmbNzvlwLDmwpi6Tbr4635w5Nw1QfzACAqg0iUUvho9mZs0gX02w7kBASKTvqd7Tyc66t1ztcuNr6YZ5524XNdZ/pQL8ZCDQj+WOdskIpZLfED3y1DQaHC7A37/C4gChXw9wbzmSru+zawhqXdE5MDgvepq3Zj24EcHMnNw6HjgX30VuywvthTSuHX5TtDrrl+8PvluGXsIsvn9f3Huj4zDYdyTuLPdYEXDfpddfc3SzDif4uxce9R/LjY878Jdrjc+fUS3PVNaAnP9a77ZAE+/mtz2POrvjVjPQa+Gvn4w1emrMPKrGz8FeVZS/YeOYEZa/bgjv+5fyEWjG2QJiLJAC5VSh1VSm1XSl2vlLpIKcWeiCWMWc3FNwvC66NldNWH8zB/8wFk2vSj2nPkhOPmI7sq9x2HjjvqW2Y14snsBB+KScv9r5LfmL4eR3Lz8cNi//4Tz00qqnH0fhrvD6vxBPvO7xsD1rWif+1xk0EOXvqT/EEHubraPjEZF4+xT5Wg74x+l2EqIbuAXz+c3+zC/NWp6wIC/Bs/W4hnJq7Ge9rFxV6bWgdjLR7gaQq84oO5IedOyi9UvvKu3X0EH87ybP/gsZN4fdr6oBO35xcU4qPZm9H1mWme1+Xk4elfVvkSduaczEe/l3/HNR85K5fxu/CyNoDFdxw4uGLRz/7ReOREy+DHTDRrMMZaTH3V3KTZdPXObFz54TxMXx1q83yRf3++EL1fnIn2o6Yg61DRMWg3LZ/XH+v24j9f/hMQ+F1rcjzlFRTi0782Ww54skvC+9Lktbj24/loPHIituw/Zhp8eWd4OZFXiDmG85pVsGasDb3wnb98iXLtGFNvOLm4NTsEX56yznakspVjhhlurPotHjx20vR7v2nvUTR7eJLfbB1mYtlfNhi7ZLYpSqkCAF2luKdpprB8+ndm1N7L2yHXTv9XnF1Zzdu8H4P+L3Ddr+ZtDUi5YcU44sl7RTgrildk01bv9vWtevLnVY5fZ9eCkK2rvdE3tQKeE9NDur5h+kTCl9nsf+MPfbOHPQOsP5y1CaN+Wulbru8QXVSDWFRYff6zH53MeGFzVtGfcd6Yvh7/MTQtZmvNb/qmTCdN1Uu0WjJvreWaXUdw4NhJFBQq335/eYp1Z3ojb0qbR39cgf+btg6zggQ4D3y/DM9ODDwWjp7Ix3eLtqPN45N9wTpQtHcPHDtpmvjWeGb2/uB7/6dOE0frTVu9x/EF00eGmvPDx/MCalSVQsBBbfa/Mr6XE8ZUIGbz8Jox/ggf0F1gGI81M96LDn0t5JSVu0xrvT6avRmjfl6FrywG1xjzBupl7i8KJv7eaD96P6TO/IZ1/9l6CC/+Zj5CPa+gMKz8kEVpQEJ+KXLzCtD5qSkBQbj+fKSn30RhoULnp6f6zVTg9cqUdSgoVHhnprOBXqH0wXOLXU2a95JgMYAJInKNiFzovcWgbESmPp+zBet2B16Fxaujrx0nAw9utmn+sGOsDRzy+p++IATw9JHzMqZK0Z+ktxhy3Xl/QJ+ZuNoyULeKh/YdPYHGIycGLXswoSYhzi9QvuBSb/HWg/hSF7R7P88UrX/TifxCdHl6Kp7+xXkQbcY7h21BoX+5V2Vl+9VQ/PCPeZPuifxCTF1lP5rxWZMch8afkBlr9mC8rtY2v1AhN68Ay0Mccfjg9/4pXpxcps/dtB8dn5yC1o/5j/RcvTMbJ7RmQRHB9oM5aPbwJMvZNOwYf/CNh+GW/TmOghWzfelUdm6e6fFp9T329uczNo8XFKqgOcT0n/ej2ZsDRo5u2ns0pMFN4Xh0/Ap0e2ZayKNdnRwz2w7kmB6bS7cfwsGcvIABYFbnnU+1OY2VUr5g2GyU7UStRt4shcy2Azn47G/v+wQve6wEnbsTQDUA++EZ0angOS8oAD+4WC6iuLFrJoyFVWEMfACA3dnOr3bfnlnUhGp29a/PEm6XRNjIrG8jYHHCVsZ1xBcg/m/+Njx/YQfH2z37zdl+j09/fjqm33smLnjHWTb7X1fsxEWGOWiNtoaR9sSsr6DVD83klf61BtsP5mC8LqjL19X22Zm1bh+Gt6/re2wMmpwwDiRywqrv6Exd4ty9R07gBW1wUVRm0wjz1zTcGvMxf2y0HRxlJ0fXVHf2m7NQMS0FczfZd0zXfzyz5kF968PwN/y/Ay9PXutLKBuKoa/PQrdGVfH0+Z55Pr2pa8wGUvyweAduPbOZ7/HtX/2Dt67sYvgM1v+j3i+a13x6a8ycphCZr40aff/PTXje4f/nu0XbMahtLeScKEDtymVx+ftzsePQcVzYpZ5vnURoQ7QL0mqKyD3wZPn3BmdeCRRnxsf9g1uFPCUOFQ+nPB7fnMkHHCQaBdw9geiv+t/SNQ0Ya8mM51+rIG3NriNB8xrNWLPH8TyOe47YN8ftPJyLi951NpLMqT4WTWmZ+475BSLR0usF/+2Z/diZHQM/LN6BWpXLRrTtcI6tCUuCD/y49YtFQZuln7Bo0jKzeOshXNw1dhdVxgDNbjcppSAi2Kd1edBfGNkNSvB7D4ufWhH7uVo37zuGcQtDzyN2OCcPq3dmY/XObDx9fjvk5hX4mt+9ScD1pq3ajdTkoga5X5btxEsXF6BcanJETYV2eSA/sDjHAOYje63MXLsHD3y3FIUKyBw93NeFolAlVoBj19yZDKCidkvX3ffeSrVh7evEuwhUyq3KCq/GzQmnJymzdCpm5m8+EJDXyDg1VrAA7Zlfiob8O6npCScVi9H3i7bbduwGgD+DpOyIh3d/3xh8JQsKyvEsBnpOmt0izYF2zzj/EYk/LN6B+77zP67cyrFnx2ywwZ9abZ3VNHlOWO2uxyestJ3zdEFmeJPWd3xqiu/+vqMnLGu6vESApwxdBbo/6xkUY+zkH4oTNqNmjU2gXs0fnuT4fOTl3b9/rtvrF3Ef9vU5jH8uRLuatJ1KqadiVpJiJgFqQamUM5uSqyTJKyj0G+L/4ezNGDt3C2Y90C9mZbj326WoXiEVix47y3IdfQXXDZ960kQseGSg78cqWn5ckoW2dSv7LQs3BYKdnBMFARcAxrk+3RKsP6NZMtqVhvyDuXnhJ3Q2OpKbh7wChWoVUm1H7hprPQEgRxvUElFYavHiJTZpYuzc/PlCjLm6q6N1uz0T/Pg1a8k8eiIfj/24wjdTzZuG9DtuzZubH+IFwMRlRSPG/UblKgQdyR5LdkEa4xAbSYnQWE3kkmsdpoBwk1nG+hP5hRjxdWDeNzcFm+fSrBN6uDN2hOpKLWdbNH1j0qG/tHbtaD/KU7OUOXo4mpoMTPlxSRZOb1bd8vVOZrCwMz9KGfq9pqzaHdaE56HSp1OZbRjxHE7qjVgzm6M5XuyCtAExK0UxxBiNSjKzLP2JIlhn63Acyc33y0dnZDd4wiwxsdM8Z6FKpPxNiSIrBk1SdrnTjHMCe3lTebSune5KmSIRz9GLZ/3fn/HbeDFk2SdNKVWy21KIiDTBrpzfcphXiWIvkvlqnTJrznRKn+uspFnocKBPvDQeOTHkhNV7j/oH/W/H+bvvZFooIiJKAPqZKqh4cDLdW6xZTTRfEv1hkmLIztDX/dPmxLupn0FamNjcSUREwYTaod1tVjMfkEc4OQLdxCAtTJwpi4iIipt41wxRaBikhSmJMRoRERG5iEFamOyyKX9+Q48YloSIiIhKIgZpLujTMiPeRSAiIqJijkFamNgljYiIiNzEIC1MjNGIiIjITQzSwsUojYiIiFzEIC1MGRXTcEWPhmhUvTwaVS8f7+IQERFRCcMgLUwigucvbI8/7u+HvhwoQERERFHGII2IiIgoATFIIyIiIkpADNKIiIiIEhCDNCIiIqIExCAtCga3re27P+bqrgCA+we3ildxiIiIqARgkBYFpzer7rs/pJ0nYPtvv+bxKg4RERGVAAzSiIiIiBIQgzQiIiKiBMQgzUVDtL5qtSuVjXNJiIiIqLhhkOai+4d4Bg/cMYD904iIiCg0DNKiQMR8tvVmGRWROXo4rjq1kd/yTg2qIDWZu56IiIisMVIgIiIiSkAM0oiIiIgSEIO0GOnbKgPXndE43sUgIiKiYoJBWox8en0PnNupbsDyR4efgkeGneJ7XKNiaiyLRURERAmKQVqUnNa0mm9KqGCU7v41pzdCrxY1fI9Z20ZEREQAkBLvApQUX998etB1zMeA+itbJjnywhAREVGxx5q0OGmaUQEAIBAoXdVa31Y1LV/z1pWd3S4WERERJQgGaTFUMc1Tcdmgajl8cdOp+OS67khN8f8XJNlUt53dIbBPGxEREZVMbO6MoRa10vHeNV3Rq3kNVEhLQb/Wnloz5ddLjYiIiIg1aTE3uG1tVEgzj40rWizXe+2yTr77k+/qE7VyERERUWJhkJYATqldCdec1gg/39Er6Lrnd66H+lXLAQDKpwYOMqhbOXAy917Na+DbW4MPbCAiIqLEwSAtASQlCZ4+vx2a1KhgOQ/ogNbmAwruOaul3+PRF3XAp9d3D1ive+NqGNK2dsDyEf2bo3alwMCOiIiI4otBWjEw+a4+eOfqLqbPjRjQwu+xCAIGI3jjPn0+NgCY9UA/3DOoFd67pisGtakVvQITERFRxBikJZiq5cv4Pf7nsbPQqnY60lKKmjbfurILhrStjbpVyvmtWyE1GR3qVUGZZPN/66XdGmBYe09tWpMaFdCgWnkAQMcGVfD+td1CLut9g1oGXWfz88McvVeK3bBWIiKiUiguQZqIvCQia0RkmYiMF5EqFusNEZG1IrJBREbGupzxUKV8KhY/dpbvcbUKgdNEdWpQBWOu6YpkQ2Cz8qkhqGwI8vRSU5Jw76BWAJwl1g1Xn5YZvvsigpa1Kvoe1zHpMwcA/+7TFI2ql3exVKFrUK1cwDKzJuN4SkvhdRYRUUkVrzP8VADtlFIdAKwD8JBxBRFJBvA2gKEA2gC4QkTaxLSUcVLVJDALhV0Alq6NIO1Qv7Lte6x7ZmjY2y9vmDVhyt1n4tyOnhxvDw5pjRt7NQl4jQBoXTs97G26QQx7cuKIXhhzjfXUXxbdCX26N64ajWKVWlVsLkCIiEqiuARpSqkpSql87eFcAPVNVusBYINSapNS6iSArwGcF6syllQ1K5XFT7f3xOiLOliu0yyjQkC/NjOnN6sesOz8TnXx+DmBsbQ+gHlwSOuA55NE8OqlnQKWh6p9PfvgMxTGFlhj0GbUqpZ9kHlH/xaWz13Ro4HjcvmVqRS1EpfjlGlEVMokQlvJDQB+NVleD8A23ePt2jIK4pQ6lVAzPc332DhitEP9KrZzhE6/t6+j7XRtVC1g2WuXdw7oKwcA7ep6gidv+hCjJIFl/rhQXN+zse3zNSoG1lKOHBoYNAKB+y2YYOvXqJhm+VzvFhmWz9lJSfL/Cg88xXpaMSIiKl5cC9JEZJqIrDC5nadb5xEA+QC+NHsLk2WWqflF5GYRWSgiC/fu3Rv5ByjGKqSlYP4jA13dRjNt7lGnbuzVBJNG9Ea3xtVMZ1ioke4fwDx9fruIymelfb3KaFGzot8yu+BJL1jMFiyka1O3Utiv9XufOkXvk5Ls/8qq5SNrKg/V65f7135mpDvbl+Fw+n8ymnI3kz4TUfHkWpCmlBqolGpncpsAACLyLwBnA7hKKWUWfG0HoG8Dqg8gy2Z77yuluimlumVkhFcrkUg+vq4bLuwcnYrDaLeIPTr8FMe1bV5JSeILUpJMop2rTm3k9/ia0xqhoTb6dNKI3ujayLo/V/nUZMx6oB9WPzXEUVnON+xX7+HX1hBEeUt5S5+mGNK2dkBwZzS8Qx1H2zcTbrNltEfF1jOpBbXTtm5lZI4e7ntcqax7M831t8gVGEy4wV20dbM5hokoMQ1vH/55PRriNbpzCIAHAZyrlMqxWG0BgBYi0kREUgFcDuCnWJUx3vq3roVXL4u8jxYQ/X5LoTYDGpVJTsLPt/vPrmAcqapXIS0ZhaZxvEfbupXQoFp5lEtNhs1qAICU5KSAJlfvS1rXNtR0aUU6pU4ljLmmK1IsUpt4Xdy1PtY/OxQ9mlTDLX2a+pZXTEvBEyb99Lwu7BJaMH57/+a++01q+NdoRjoL7PR7z4zo9a9d1jnCEkRfonTbs+tiQMWLWbcJIjfEq0/aWwDSAUwVkSUiMgYARKSuiEwCAG1gwe0AJgNYDWCcUmplnMpLOvofvaHtwktJ0T7I6FLAf+J5s+Dr2Qs8TaKFFpGJPpUJANzcpymeu6A9zu1YF1/edCo6N/TP/GIVe5o1z1opk5yEcbecjoeGneJbNu2eM3F9T8+IVmNNVZ3KZfHKJR1NaxfNZI4ejmG6K7v/9mtus3Ys+O+bdvWsm3Qj5WQXfXPzaQHLkpOd7ds/7u8bYolC4/YgD7va5ngzm8KuOPv5jl4Jvb/dFGkaoi9vOjVKJfGoU7ksJvy3p+XzxT2gjtfozuZKqQZKqU7a7VZteZZSaphuvUlKqZZKqWZKqWfjUVYCejTxDBC4V5uCSp8H7d2ru/o1d3n9fHsvfHJd4PRUofAGZgKBWYu4d0J6s1q2CzvXC0hl8vCwU5CRngYRQc/mNTDultOx+qkhvibDlCRBY12utl7NPTM0tKjpLDWI1W9wbV1uuB9uO8PvuZFDW0NE0L91zaCDHswYg7tgNYmRClb1H24ta8cGVdAxSOBuNiDFKCM9LaCbQKWyzlJ3NKoeWj/LUAUb/Rupcbck7vy8/xgumIq7OpXLhd38Himz/I3R0LSGs+O/a6Oq6G2YvSYUPZuH/1oz3//nDFS3CcRm3Nc3ovcP5SLdDYkwupMS3LhbTkfm6OG4Y0ALZI4ejuZB+mYBnpqyfhGexHxBmgAFJtGHNyDQ16QZ15p2j3XzXZnkJJRLTcbZHerixl5N8NDQU/yCjIu71sfCRweinSGth9kk9mY61K+MK09t6LesVqWyuGtgC9Sq5Okn5U1Gm5KchCfOaevofe14TyjGPHjNa1bE/YNbBX19WkoSrjmtUcCAAK9bz2wWcRmNKqQm452ruthWNY25ugsu6WqWqSfQy5d09N2/QAvYnP7P3NS7ZQYWPWo+oCfUgThmr7frMhAKNwbtlLam3kj6pwbTtEbw829pImJ/cVjcU/cwSCsFEqVPjpH3B3/9s0WJc/u0zMBT5wUGK4WFga/3/iaZjjvRnnMSUKamJOGxs9ugcvkyfu9VPjXFtNP57Af740eb6nWvn27vhecuaB+w/K6BLdGxvukkG368NZfhuPb0xn6PP7+hh6OmURHB0+e3c1Q+IDo1dyMGtEC9KuUCjlP9/65PywxHtXQigiRdsOIN2P5+aEBETbH1qpTD/13W0W9ZelqK5QwaRp/f0ANntsxApXJFtXo36/otNongh3fM1V1tL0ZCdc1pjYKvFILTmwbmUywJzMe7ebx9ZeBcy8Yp/8JVmnIjRmrRowMtp0l0qnK5+CbRZpBGPrE+mY4c2hqZo4f7fYk+v6FHQIAhYt4h3ptcVt/c6R2BeWqTwBxuoXh4WGvLAC8pSdCpgUkQE8LJs6jENleAIfbjmXlfX8uRA96mwvM61Q3pPY3caGox+9F55vx2YQUexrfS1y6ZBfpXn+ap6Xx0uKcPodXo1jeu6IQLOtfHLWcWBVbdbGaQGH2hf3Cu7yLgVdNBuhJj7adxoIhXpIN5wnX3wOAXEmNv7AHA+jvp5ohgMy9cFHjhFA2D29bye/zSxf4Jw4Ndz7xnM5tJIvA2h3Z1MHPKv3sHzioTD9UjHNl954AWeHR4fCc6YpBWgn2oTZru9AT+2Q09sHzUIDeLFJLRF7VHmzqVUKtSWRSajA7wfiz9j2/HBlUw7+EBuLRbeBn8L+jsaVK7rFvDIGsGCjYjgZ6+KdeK8QrQbF39XmlSo4LvsX7VeQ8P8N132unX7AelYbXyqBLjPGznaNOJeZP2ejugh9MfSB/M/7t3E3x2Qw+MOqctPr6um2+qsk+v9/SjzEhP85t1w/vSDvU8wXnTGhXw9lVdLGsSrcqn/794B5NYuX9wq4DaT7Pam0rlzIOcS7rWd9zPKFze5vW+rTJw18AWuO6MxgCKglIR+EZFh9Ic27O5exeMZgHzRV3Mm9L7tXKezsk4z7JxYJLZOUzvNN1F8os2M8KEUnvttB9kelpK0D6fvVvUwLpnhqJLw6pByzDwlFr2K2iWjxqEX+/s7WhdKwJBLZfyM97Yu0lUkqxHgkFaKeD01JiakoR0h52sjc5qUwtly0T3cOrdIgOT7uyNMslJpoMDvOd843O1KpUNu2ZhxIDmWPO0/UT10eEps7GU6WkpGNy2Fm7r2wxX9PAPFJ18osZa5/ealYpOWrUqFTXJOT2/e4OBxkEmvY9Gl1pvcKvP+eZ931cu6Yg5D/X3BUyzH+yPP+7v69sXDxlmi3D6bx/ctjbObJmBlOQk9G9dy3e8eI9/gX8NT1qKJzjs2MDT1+/Rs09B+VTrk3dGelrAj7SRPmgxK7f+/2bloaGtcUazoo7YmaOHo68WWAxrXyfiTtPB+ILX+lVw18CWvg7l3uPH6ahloy9vChyl6zWif3NfMG3UslZR7XftSmVNa8OVAl4zpDd65dKOAesBwNAIcmQZZwOpU9lZLXR62RR0MRk56m0lcPqd+/Dabn79aY0DS/Tf7Ul39sZbVwZPn2OcLvCVS8z3m9Pzb3rZMjilTmQjwvUXAlacdksIeO+wXhVdDNJKsAZaMtgeETb9OfHBtd2w5unwJ2UPxuyL7D0RBLtCn37vmZivq02yIyJhd3IO5/fIeDJb/uRgvHdNNzwwpDVSU5Iw4b898YxNR+6ehvlT/9uvGT6/oUfY00wFK59RsCvqTyx+TP234fnr94OovXFqSpLfj1u1Cql+ozCbZTjvy6VPYhysn4qI54cL8DR5eFPG1K9aHpmjh6N/a09NgdnIr6WPD4KIYPxt1v0WndQqma1hbL65xWQgh9X/5AaTmjurAQsDWtc07VdlxVtW4+Gif2h2KBlrn5xoU7eS5dzC9+j6cZ5SJx0faK0JRsaE1gDw5/398Of9/fyWhVIjb9zvFQ3NuI8MPwVG/+nbDEsePwu/23RV8PJeeJnVpj5q8t76GqBHhp0S8Dtwt25fNQixltx73OsvBqNlUBvzWrgzTWpAnZrzkLPzv1G8uhHoMUgrAS6wmJmgVe10zHqgn18H5eLqhYs6BOTX6dqoKm49sxnGXG3fl6NZRkXUdFArEalQvs5Omyw6NqgSUKPm1bRGhYAryJTkJNPmnFC3a7ZaOLmu+rVy3ix5gy4FySVBfhwfHNoanRpUwemGINWuyfmWPk0x7+EBeOKcNgGjX83UTC+LzNHD/X7MjMz2Z7RqYY1v/eJFHTDm6q745Y5epusHMOyKISY5Da2CpI+u6x71EYrt6xXVLK54cjBeuKg9xhtS0jgxpJ11ufT/j1ObVrfsw6fnzaPVsHp5NAxSc3x596Lj0vgD7t22d3SsftBR5XJlAgLLiSN64YHBrVClfCoa68opsL/gMwscjCkoRgxogdOaFgVloR6TdheGfmWx+L410u1HpxfIQbdlsU+8FzyOvxehbDPq7xg6BmklwP9d1sk0VxnguUJKhKuBSJVLTfbl16lftRxmPdAPNSqmYeTQ1r4aw1irYAhaQum74K2ZcVIN7610uaN/C9+yiSN6BeRcc8I41ydgPkS9FI1elwAAEo5JREFUKEddUSf4Zlpzy9c3n+bLUWfGm15kzNXmNTH6Hzo9EcHm54dh8/PDgtZmtqyVjh//2zNgnwf7catVqSyu79nE8jsRak6kWGRQ+vDabvjljl64tHsDZKSnBaSEMfKOzmxrqH02fuSa6WlBg3armigvq5cnJwn+dXojjLu1qIntvkFFwW7FtBRc1r0hGlWvgMfOLuqYPaJ/c0f96Lypa4yd9b3laV+vst+sH3rGZmT99vUm3xU45+tzF7THumfsWwzKmNSSKqUC9nVKUpLfcehtHc1IT/PNimJWQ9beZJSy8b1v6u05xgNqNnWPzWrOvAHrYJO+q+V13zW742bz88P89nHNSmV9TfBeLWtVxPkhDmIyftfba98DbzAc7HsRjHFaQCAxRtIySKNiZdGjAzHl7j5xC8z0vtH6eNSp7KlxCaWZ9I7+LTBxRC9HJxYRCajRaVu3clid+AeeUgu392vua1IY0b85/h7Z32bjgSfH05pWR90qnpOwWVDjPXkmJwWeXlKTkzDaplO054clfmdGfQLlUHhHMIZLvzXj4I6BbWqF9AM0sE0tZI4eHrT2uHrFtKBB5lkWTU+D2tRCu3qVUE2rofHWyKUme74D6WXL4Mnz2qFLw6K+VVb9hvS1pnef1dJRP7ouDavi8bPb4AXDseT9/9WvWs7yODI2NadalKtV7cBO90lJ4qsRMzY72gX4Tmqw08uWwcuXdMSXN52GtJRkZI4ejpt6FwWa3mOyUTVPEKu/wMswdJy3Onq9y9+8ojP6hJiQdoTuItH3fiYbMtvvn17v//2YcveZeO3yoj5w+gs6q6//s+e38wv0x992BlY9Ndh03XBmgzA7f4d6HnBDfIctEIUo0iHV0eRN1+Ck6cwoOUnQtm5kV35OpBv6xSQnCe4b3Ar3f7sUgKePlXFmBo+iXxVvzrTzOxU1qzs5eZn1nfH+QL5/TVfcPHaR570S4XLVwGmRvB8xktkE/tO3GTbuOQoAeOeqLpi2enfY7xUK/5yAycg5WeD4te9rNWwFhQrl01J8oyN7Nq+OkUNb44ruoY+ODoWI4IZe/n3s1j87FJNX7gLgbMBCxwZVsHTbIcf/62D9K2trfSfNvk8Kgd9FMxebJGx+/fJOSBJBjYppmJ95wHRKvd4tMvDp9d1x6xeLkJtnkmvGYFj7Oqbfu8u6N8DbMzea1pTrUwJd1KU+/t64H80yKqJsmST0ap4R0XGrb8I2DrjwqlI+Fbf3b4EaFdN8gwWsAv/v/+OsleGKHg3wv/nbLJ9PhFMTgzSydEWPhr5aEQpUtUIqJvy3J1rUSswM4K9f3gmdG5hfUQa7sK9WwRMMD2tXBw2rlw9oTr9rYAv858t/0KBqed1rvD9O5me2Owe08PWLGtS2Nq47ozE+/Tsz6OcIRaQn1bCbLwWYencf5AdJs2Dk3a+3jF0IwNO0Xdf7Y2/Tj6h746rYlZ0bZmGLPHZ2GzwyfjnGXN0VvV+cGfLrk5PEr3O9iIQ1K8WI/s3xxowNIb9Or0xyEhpqNez6kbXjbzsDPy/diY//2uy3/lWnNsTSbYfQ3OG0b8H6V97Rvzla1qpo2vG9WUYFtKtXGR9e2w1P/rIS2w4cd7RNADhPd3G07pmhyDpU9NqZ9/VF9vE8AEDfVjWRrH0BrC58WtRMx9rdRyy3dd+gVrhzQEvLgRleF3Wtj4u0gNI7YKzxyIkOPk1wo85ti2oVUjF27hYAwC1nNsWUlUUB4OUWfXRDtem5YRABsnPzMaxdHdSunIb7v12GTfuOReX9o4VBGll6/kJ3kj6WJB3NktomCP3JPVTVKqRi6eODLK/+h7av4xe4zbyvry+oGNy2Fqat3o0WhtolYwd8b7+V6mGM8HOLt3YphIx3vnvGz+ulb11zEsLdObAF2tStZJsL7ttbQ++PaKZTgyqYOMI+T9W0e87Ehe/8hezc/Khs08w9g1rhnkHW05a9dlkn3PXNkqDv06F+Fcy490y/AQOdG1ZF54ZVA4K0S7s1wLD2dQJqje49q2VYzWVlkpNwdgf/flZVypfBoZw8vHmFpzlvYJtaeOG3NQDCv6CoUdHTZ+2Jc9oGDIz44Npu+Gj2ZpS36Hrx1b9Pxaqd2Zaji0UEqSnRqT7q3LCKL29eKDLS0/D0+e3QrXFVTF+9Bw8NPQUPDQ3smxcp78wk+lHMM+7ri9nr9+Hqj+YBYE0aEcWJ8dzz5hWdkZvn39wVyogw/Y/FJd0a4JyOdYP20ft37yaoV7UczoniKMJoNZ2G+j5Wzb/jbjnd13/Pf31rZZKTMCyC/FxOOU0v0bxmRXx6Qw+8PWMDpq/Z43Kpiqx6ajDaPD4ZgCdlhpMgDQCahpCWxaxZ744BgX2vwuX9P1dIi978kakpSZj9oHk/0jOa18AZNhOYV6+YFrX0PMHYpaBx4rxO9SK60AxXrxY1kCSeOaHZJ42I4sJYo+PN7B8t+gBt/sMDTJsBU5KTcG6Ut5scYZDmnVtzqEm6CjPBOoSHmqMwGnOhWtHvmc3PDwvptV0aVsVH13WPWpOWE+VTU1C1fBnLwQvFweU9GuLd3zf6Db65vEdDPP3LKtRKdz8t0J0DWmDbgRzT9CtO/ev0RlFpWi9OfDO3xD9GY5BGVKrF4CTkdo66Ef2bY/LK3bj6tIaorY14e/eqLqaj84KpVLYMljx+luOZN6J1Mo/1FXsiDtYws/jxomnqburVxDRgm/VAv4BlVqI5WbZ+xO2bV5hn639gcCvcNbCFb8YKALixVxPfNGRua1CtvG8Uupmfbu8ZNKHwk+c5y5mWCO4c0CIg3Uc49CmI4o1BGlEp5GaNTayZ9WeKZDqfkDKvh9iHLSVJ0Lp2esCcnN4p1UKZ3zJUidx/0olHLfKZOU3HM+uBfo5GWBrNf3iA6T+4b6uaqF2pLHZl51rWRIuIX4CWaDrUj90x8dtdvVHBZiq1aLBLPB2OcKc1iyYGaUSlWPxPQcVbUU2asz0pIvjNJEnqqHPbon7V8hjgcGLqcASbCqukCze3ol1N8LR7z8TuUtYUGK7WtSObozMeEiBGYzJbotIo1Mz6ZC/Sc3mV8qm4b3ArV2vSAPM8XBS+imkpIc0fS8VLInQLYE0aUSmWCCeh4uyx4W3wyI/LAybTTlQvX9IRL1/SMW7b/+bm00ynJiN64pw2ttPNRcvgtrWwIPOg7TqNqpfHlv05rpfFieJxZiEiSx3qV8ay7YfjXYxSSZ/Uk4I7tWn1eBeBEtT1PWMzmOK9a+znowWAb285HSuzsmNQmuAYpBEVc+Nv62k6BZOdO/q3wJqdR3CWi32giIiKo5qVyro+Kt0pBmlExZynH1NoTUhNalTApDvtM80TUemWHoPmR7LH/wARUQm24JGBAbNJRKJ1GPnnqPhZPmpQSANZHh7WGh1jmNKjtGCQRkRUgmWkp0XtvdY+MyTiWR2oeHCa0Nnr5j7NXCpJ6cYgjYiIMPehAUhNsc/KlMiJWYmCcTvFjRsYpBERkW9KLaKSaP4jA5BaDBM6M0gjIiKiEq1mDCa0d0PxCyuJiIiISgEGaUREREQJiEEaERERUQJikEZERESUgBikERERESUgBmlERERECYhBGhEREVECYpBGRERElIAYpBERERElIAZpRERERAlIlFLxLkPUicheAFtc3kwNAPtc3kZxxX1jjfvGHvePNe4ba9w31rhv7CXK/mmklMowLiyRQVosiMhCpVS3eJcjEXHfWOO+scf9Y437xhr3jTXuG3uJvn/Y3ElERESUgBikERERESUgBmnhez/eBUhg3DfWuG/scf9Y476xxn1jjfvGXkLvH/ZJIyIiIkpArEkjIiIiSkAM0kIkIkNEZK2IbBCRkfEuT6yISKaILBeRJSKyUFtWTUSmish67W9V3foPaftorYgM1i3vqr3PBhF5Q0QkHp8nUiLysYjsEZEVumVR2x8ikiYi32jL54lI41h+vkhY7JtRIrJDO36WiMgw3XOlad80EJGZIrJaRFaKyJ3a8lJ/7Njsm1J/7IhIWRGZLyJLtX3zpLacx431vikZx41SijeHNwDJADYCaAogFcBSAG3iXa4YffZMADUMy14EMFK7PxLAC9r9Ntq+SQPQRNtnydpz8wGcDkAA/ApgaLw/W5j7ow+ALgBWuLE/ANwGYIx2/3IA38T7M0e4b0YBuM9k3dK2b+oA6KLdTwewTtsHpf7Ysdk3pf7Y0T5HRe1+GQDzAJzG48Z235SI44Y1aaHpAWCDUmqTUuokgK8BnBfnMsXTeQA+0+5/BuB83fKvlVInlFKbAWwA0ENE6gCopJSaozxH++e61xQrSqk/ARwwLI7m/tC/13cABniv6hKdxb6xUtr2zU6l1D/a/SMAVgOoBx47dvvGSmnaN0opdVR7WEa7KfC4sds3VorVvmGQFpp6ALbpHm+H/UmkJFEApojIIhG5WVtWSym1E/CcYAHU1JZb7ad62n3j8pIimvvD9xqlVD6AwwCqu1by2LhdRJaJpznU2yxTaveN1mTSGZ4rfx47OoZ9A/DYgYgki8gSAHsATFVK8bjRWOwboAQcNwzSQmMWOZeW4bE9lVJdAAwF8F8R6WOzrtV+Kq37L5z9UdL21bsAmgHoBGAngFe05aVy34hIRQDfA7hLKZVtt6rJshK9f0z2DY8dAEqpAqVUJwD14an5aWezOvdNCTluGKSFZjuABrrH9QFkxaksMaWUytL+7gEwHp6m391aFTG0v3u01a3203btvnF5SRHN/eF7jYikAKgM502ICUcptVs7kRYC+ACe4wcohftGRMrAE4R8qZT6QVvMYwfm+4bHjj+l1CH8f3t3F2JlEcdx/PujiA0Ta8vClsCVlkzLlbZAYXfpQhaRiDLJLkJBIeomCiIvJCEQNJSuypsyogIvIlaFIoWiVlBRTPfFNSsrL4q86kVbCtN/FzOnHg9nN6utfc45vw88nOF5mZlnmD3735nn2YGPgKW431yi2DaN0m8cpP09h4EOSe2SriI9QLh7iuv0n5M0TdL0ShroA0ZI9746n7Ya2JXTu4FH8hsx7UAHcCgPx5+VtCjP568qXNMIJrM9inmtAD7Mz0nUpcovkuxBUv+BJmubfC/bgRMR8WLhUNP3nfHaxn0HJM2UdG1OXw0sAT7F/WbctmmYfvNP3zho1g1YRnrr6BSwfqrr8z/d8xzS2zCDwPHKfZPm5D8APs+frYVr1uc2OknhDU7gbtIPyyngJfI/VK63DdhBGkI/T/ora+1ktgfQArxNeqj1EDBnqu/5X7bNm8AwMET6wpvVpG3TTZomGQKO5W2Z+86EbdP0fQdYABzNbTACbMj73W/Gb5uG6DdeccDMzMyshDzdaWZmZlZCDtLMzMzMSshBmpmZmVkJOUgzMzMzKyEHaWZmZmYl5CDNzMzMrIQcpJlZ05H0taQbJiGf1yWtmIw6mZlVc5BmZnVNib/LzKzh+IvNzOqOpNmSTkjaBnwCPCfpsKQhSc8Xztsp6Yik45Ieq5HPNEnvShqUNCJp5QRlbpY0msvYWjjUK2m/pC8ro2o5cNyS8xyu5Ctpm6T7c7pf0ms5vVbSxsJ9vZLrvDcvdWNmTchBmpnVq9uAN4B1QBtpAeWFQJek3nzOmojoIi338qSk66vyWAp8GxGdEXEH8H6tgiS1ktb/mx8RC4CNhcOzSEsa3QdszvuW57p0ktYS3JLXEhwAevI5bcC8nO4G9uV0B/ByRMwHfgAeurzmMLNG4yDNzOrV6Yg4CPTl7ShpVG0uKdCBFJgNAgeBWwr7K4aBJZJekNQTET+OU9ZPwC/Aq5KWA2OFYzsj4mJEjAI35X3dwI6IuBARZ4CPgXtIgViPpHnAKHAmB2+Lgf352q8i4lhOHwFmX36TmFkjcZBmZvXq5/wpYFNELMzbrRGxXdK9pFGsxRHRSQriWooZRMRnQBcpWNskaUOtgiLiN9JI3TvAA1w64vZrIa2qz+p8vgGuI43gDZCCtoeBcxFxtkZ+F4Ara9++mTU6B2lmVu/2AGskXQMgqU3SjcAM4PuIGJM0F1hUfaGkm4GxiHgL2ArcVauAnPeMiHgPeIo0lTmRAWClpCskzQR6gUP52IGcRyVIe4Y/pzrNzP7gv9DMrK5FxF5JtwMHJAGcAx4ljXY9LmkIOEma8qx2J+l5sYvAeeCJcYqZDuyS1EIaJXv6L6rVT5rCHAQCeDYivsvH9gF9EfGFpNNAKw7SzKwGRcRU18HMzMzMqni608zMzKyEPN1pZlYgqR9or9q9LiL2TEV9zKx5ebrTzMzMrIQ83WlmZmZWQg7SzMzMzErIQZqZmZlZCTlIMzMzMyshB2lmZmZmJfQ7wA4WY5HZLvcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import math\n",
    "current_date=datetime.datetime.today().strftime('%Y-%m-%d')\n",
    "\n",
    "# Accuracy\n",
    "plt.figure(figsize=(10,5))\n",
    "plt.title(\"SSIM\")\n",
    "plt.plot(epochs_list, ssim_average[\"training\"], label='training')\n",
    "plt.plot(epochs_list, ssim_average[\"validation\"] , label='validation')\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"SSIM\")\n",
    "plt.legend()\n",
    "if not flag_debug:\n",
    "    plt.savefig(os.path.join(output_save_directory, current_date + \"_accuracy\"+\".png\"))\n",
    "    \n",
    "# All losses\n",
    "plt.figure(figsize=(10,5))\n",
    "plt.title(\"Losses\")\n",
    "plt.plot(epochs_list, loss_per_epoch[\"training\"], label='training')\n",
    "plt.plot(epochs_list, loss_per_epoch[\"validation\"] , label='validation')\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "if not flag_debug:\n",
    "    plt.savefig(os.path.join(output_save_directory, current_date + \"_loss\"+\".png\"))\n",
    "\n",
    "# Loss for training\n",
    "plt.figure(figsize=(10,5))\n",
    "plt.title(\"Loss during training\")\n",
    "plt.plot(reals_shown, [math.log10(y) for y in training_loss_list])\n",
    "plt.xlabel(\"reals_shown\")\n",
    "plt.ylabel(\"Training Log10(Loss)\")\n",
    "if not flag_debug:\n",
    "    plt.savefig(os.path.join(output_save_directory, current_date + \"_training_loss\"+\".png\"))\n",
    "\n",
    "\n",
    "    \n",
    "# Final Model:\n",
    "with torch.no_grad():\n",
    "    input_image = fixed_val_sample['source']\n",
    "    input_images = vutils.make_grid(input_image[0:1,:,:,:], padding=2, normalize=True)\n",
    "    target_images = vutils.make_grid(fixed_val_sample['boneless'][0:1,:,:,:], padding=2, normalize=True)\n",
    "    net = net.cpu()\n",
    "    output_image = net(input_image[0:1,:,:,:]).detach().cpu()\n",
    "    output_images = vutils.make_grid(output_image, padding=2, normalize=True)\n",
    "print(str(torch.max(output_images)) + \",\" + str(torch.min(output_images)))\n",
    "plt.figure(1)\n",
    "fig, ax = plt.subplots(1,3, figsize=(15,15))\n",
    "ax[0].imshow(np.transpose(input_images, (1,2,0)), vmin=0, vmax=1)\n",
    "ax[0].set_title(\"Source\")\n",
    "ax[0].axis(\"off\")\n",
    "ax[1].imshow(np.transpose(output_images, (1,2,0)), vmin=0, vmax=1)\n",
    "ax[1].set_title(\"Suppressed\")\n",
    "ax[1].axis(\"off\")\n",
    "ax[2].imshow(np.transpose(target_images, (1,2,0)), vmin=0, vmax=1)\n",
    "ax[2].set_title(\"Ideal Bone-suppressed\")\n",
    "ax[2].axis(\"off\")\n",
    "plt.show\n",
    "if not flag_debug:\n",
    "    plt.savefig(os.path.join(output_save_directory, current_date + \"_validation_ComparisonImages\"+\".png\"))\n",
    "\n",
    "# ANIMATED VALIDATION IMAGE\n",
    "fig = plt.figure(figsize=(8,8))\n",
    "ax = fig.add_subplot(111)\n",
    "plt.axis(\"off\")\n",
    "ims = []\n",
    "training_ims_shown = []\n",
    "for i, im in enumerate(img_list):\n",
    "    if i % 50 == 0:  # controls how many images are printed into the animation\n",
    "        training_ims_shown = i*(100*_batch_size)\n",
    "        frame = ax.imshow(np.transpose(im,(1,2,0)))\n",
    "        t = ax.annotate(\"Reals shown: {}\".format(training_ims_shown), (0.5,1.02), xycoords=\"axes fraction\")\n",
    "        ims.append([frame, t])\n",
    "ani = animation.ArtistAnimation(fig, ims, interval=200, repeat_delay=1000, blit=True)\n",
    "if not flag_debug:\n",
    "    ani.save(os.path.join(output_save_directory, current_date+\"_animation.mp4\"), dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c923046b",
   "metadata": {},
   "outputs": [],
   "source": [
    "clean_matrix = torch.from_numpy(np.random.rand(15,1,256,256)).float()\n",
    "true_matrix = torch.from_numpy(np.random.rand(15,1,256,256)).float()\n",
    "print(criterion_Rajaraman(clean_matrix, true_matrix, alpha=0.84))\n",
    "print(len(true_matrix))\n",
    "print(len(ds_training))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ccffbd6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "ssim_average[\"validation\"]"
   ]
  }
 ],
 "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": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}