Nitish-Gaddam-CIFAR-10.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Jh_ZkVCbGsI2"
   },
   "source": [
    "# Instructions\n",
    "\n",
    "This is an imperfect implementation of ResNet configured to train with the CIFAR-10 dataset.\n",
    "\n",
    "There are a few bugs and it should be able to reach a higher accuracy. See what you can do to fix it and write a few paragraphs about what you did."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4iI079DBDSW9"
   },
   "source": [
    "# Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 51,
     "output_extras": [
      {}
     ]
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 444,
     "status": "ok",
     "timestamp": 1522769563590,
     "user": {
      "displayName": "Jim Fleming",
      "photoUrl": "//lh5.googleusercontent.com/-bMUyWrg0nTk/AAAAAAAAAAI/AAAAAAAAADA/bxg68Muit30/s50-c-k-no/photo.jpg",
      "userId": "110726001486577440958"
     },
     "user_tz": 420
    },
    "id": "2iVqm5Jr-078",
    "outputId": "7a1fe860-7553-4e82-e14f-c64c012c8d6e"
   },
   "outputs": [
    {
     "ename": "ImportError",
     "evalue": "cannot import name pywrap_tensorflow",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mImportError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-14-d1e32a28a86e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtarfile\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0;32mimport\u001b[0m \u001b[0mtensorflow\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mtf\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mmatplotlib\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpyplot\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mnumpy\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/Users/gaddamnitish/anaconda3/envs/py27/lib/python2.7/site-packages/tensorflow/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     22\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     23\u001b[0m \u001b[0;31m# pylint: disable=wildcard-import\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 24\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpython\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m  \u001b[0;31m# pylint: disable=redefined-builtin\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     25\u001b[0m \u001b[0;31m# pylint: enable=wildcard-import\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     26\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/Users/gaddamnitish/anaconda3/envs/py27/lib/python2.7/site-packages/tensorflow/python/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     47\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mnumpy\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     48\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 49\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpython\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mpywrap_tensorflow\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     50\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     51\u001b[0m \u001b[0;31m# Protocol buffers\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mImportError\u001b[0m: cannot import name pywrap_tensorflow"
     ]
    }
   ],
   "source": [
    "from __future__ import absolute_import\n",
    "from __future__ import print_function\n",
    "from __future__ import division\n",
    "\n",
    "import os\n",
    "import time\n",
    "import tarfile\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "from google.colab import files\n",
    "\n",
    "print('TensorFlow Version:', tf.VERSION)\n",
    "print('TensorFlow GPU:', tf.test.gpu_device_name())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "JKaCOUwrK1au"
   },
   "source": [
    "# Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 34,
     "output_extras": [
      {}
     ]
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 269,
     "status": "ok",
     "timestamp": 1522769564675,
     "user": {
      "displayName": "Jim Fleming",
      "photoUrl": "//lh5.googleusercontent.com/-bMUyWrg0nTk/AAAAAAAAAAI/AAAAAAAAADA/bxg68Muit30/s50-c-k-no/photo.jpg",
      "userId": "110726001486577440958"
     },
     "user_tz": 420
    },
    "id": "7gy8inTwL8lI",
    "outputId": "3a1287b8-8f81-4984-b28e-89c62327a6b9"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/content/.keras/datasets/cifar-10-batches-bin'"
      ]
     },
     "execution_count": 34,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_name = 'cifar-10-batches-bin'\n",
    "data_dir = os.path.join('/content/.keras/datasets', data_name)\n",
    "data_url = 'https://www.cs.toronto.edu/~kriz/cifar-10-binary.tar.gz'\n",
    "tf.keras.utils.get_file(data_name, data_url, untar=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "H_xuw_3zKwhs"
   },
   "source": [
    "# Inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "DRE_SKGgKzQb"
   },
   "outputs": [],
   "source": [
    "def get_input_fn(filenames, batch_size, num_epochs, shuffle=False, augment=False):\n",
    "    label_bytes = 1\n",
    "    image_height = 32\n",
    "    image_width = 32\n",
    "    image_depth = 3\n",
    "    image_bytes = image_height * image_width * image_depth\n",
    "    record_bytes = label_bytes + image_bytes\n",
    "\n",
    "    def _parse_record(value):\n",
    "        raw_data = tf.decode_raw(value, tf.uint8)\n",
    "        raw_data.set_shape([record_bytes])\n",
    "        label = tf.squeeze(tf.cast(raw_data[:label_bytes], tf.int32), axis=0)\n",
    "        depth_major = tf.reshape(\n",
    "            raw_data[label_bytes:],\n",
    "            [image_depth, image_height, image_width])\n",
    "        image = tf.transpose(depth_major, [1, 2, 0])\n",
    "        image = tf.to_float(image) / 255\n",
    "        if augment:\n",
    "            image = tf.image.random_flip_left_right(image)\n",
    "            image = tf.image.random_brightness(image, max_delta=0.2)\n",
    "            image = tf.image.random_contrast(image, lower=0.2, upper=1.8)\n",
    "            image = tf.clip_by_value(image, 0.0, 1.0)\n",
    "        features = {'image': image}\n",
    "        labels = {'label': label}\n",
    "        return features, labels\n",
    "        \n",
    "    def _input_fn():\n",
    "        with tf.device('/cpu:0'):\n",
    "            dataset = tf.data.FixedLengthRecordDataset(filenames, record_bytes)\n",
    "            dataset = dataset.map(\n",
    "                map_func=_parse_record,\n",
    "                num_parallel_calls=4)\n",
    "            if shuffle:\n",
    "                dataset = dataset.shuffle(buffer_size=10000)\n",
    "            dataset = dataset.batch(batch_size)\n",
    "            dataset = dataset.prefetch(buffer_size=1000)\n",
    "            dataset = dataset.repeat(num_epochs)\n",
    "            iterator = dataset.make_one_shot_iterator()\n",
    "            return iterator.get_next()\n",
    "    return _input_fn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 598,
     "output_extras": [
      {},
      {}
     ]
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 4056,
     "status": "ok",
     "timestamp": 1522769569409,
     "user": {
      "displayName": "Jim Fleming",
      "photoUrl": "//lh5.googleusercontent.com/-bMUyWrg0nTk/AAAAAAAAAAI/AAAAAAAAADA/bxg68Muit30/s50-c-k-no/photo.jpg",
      "userId": "110726001486577440958"
     },
     "user_tz": 420
    },
    "id": "GbF8GxfvMs5K",
    "outputId": "fad015ee-3009-419a-a0db-7575cbbe85ee"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Images Min: 0.0 Max: 1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiUAAAI0CAYAAADC51nVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1e4a33d950>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    input_fn = get_input_fn(\n",
    "        filenames=[os.path.join(data_dir, 'test_batch.bin')],\n",
    "        batch_size=16,\n",
    "        num_epochs=1,\n",
    "        shuffle=False,\n",
    "        augment=False)\n",
    "\n",
    "    features, labels = input_fn()\n",
    "    image_batch, label_batch = sess.run([features['image'], labels['label']])\n",
    "    print('Images Min:', image_batch.min(), 'Max:', image_batch.max())\n",
    "\n",
    "    fig, axes = plt.subplots(4, 4, figsize=(8, 8))\n",
    "    for i, ax in enumerate(axes.ravel()):\n",
    "        ax.imshow(image_batch[i])\n",
    "        ax.grid(False)\n",
    "        ax.axis('off')\n",
    "        ax.set_title('Label: {}'.format(label_batch[i]))\n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "v9ih6KgrLCgf"
   },
   "source": [
    "# Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "RbGhOg7vkwPo"
   },
   "outputs": [],
   "source": [
    "def batch_norm(inputs, momentum, training):\n",
    "    return tf.layers.batch_normalization(\n",
    "        inputs=inputs,\n",
    "        scale=False,\n",
    "        momentum=momentum,\n",
    "        training=training,\n",
    "        fused=True)\n",
    "\n",
    "def resnet_block(inputs, filters, strides, training, params):\n",
    "    kernel_initializer = tf.initializers.variance_scaling(scale=2.0)\n",
    "    kernel_regularizer = tf.contrib.layers.l2_regularizer(scale=params['l2_scale'])\n",
    "\n",
    "    inputs = batch_norm(\n",
    "        inputs=inputs,\n",
    "        momentum=params['bn_momentum'],\n",
    "        training=training)\n",
    "    inputs = tf.nn.relu(inputs)\n",
    "\n",
    "    hidden = tf.layers.conv2d(\n",
    "        inputs=inputs,\n",
    "        filters=filters,\n",
    "        kernel_size=3,\n",
    "        strides=strides,\n",
    "        use_bias=False,\n",
    "        kernel_initializer=kernel_initializer,\n",
    "        kernel_regularizer=kernel_regularizer,\n",
    "        padding='same')\n",
    "\n",
    "    hidden = batch_norm(\n",
    "        inputs=hidden,\n",
    "        momentum=params['bn_momentum'],\n",
    "        training=training)\n",
    "    hidden = tf.nn.relu(hidden)\n",
    "    hidden = tf.layers.conv2d(\n",
    "        inputs=hidden,\n",
    "        filters=filters,\n",
    "        kernel_size=3,\n",
    "        strides=1,\n",
    "        use_bias=False,\n",
    "        kernel_initializer=kernel_initializer,\n",
    "        kernel_regularizer=kernel_regularizer,\n",
    "        padding='same')\n",
    "\n",
    "    inputs_padded = tf.layers.conv2d(\n",
    "        inputs=inputs,\n",
    "        filters=filters,\n",
    "        kernel_size=1,\n",
    "        strides=strides,\n",
    "        use_bias=False,\n",
    "        padding='same',\n",
    "        kernel_initializer=kernel_initializer,\n",
    "        kernel_regularizer=kernel_regularizer)\n",
    "    hidden = hidden + inputs_padded\n",
    "    return hidden\n",
    "\n",
    "def resnet_layer(inputs, filters, strides, training, params):\n",
    "    hidden = resnet_block(inputs, filters, strides, training, params)\n",
    "    hidden = resnet_block(hidden, filters, 1, training, params)\n",
    "    hidden = resnet_block(hidden, filters, 1, training, params)\n",
    "    hidden = resnet_block(hidden, filters, 1, training, params)\n",
    "    hidden = resnet_block(hidden, filters, 1, training, params)\n",
    "    hidden = resnet_block(hidden, filters, 1, training, params)\n",
    "    return hidden"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "Ch-hDue_LEu-"
   },
   "outputs": [],
   "source": [
    "def model_fn(features, labels, mode, params):\n",
    "    kernel_initializer = tf.initializers.variance_scaling(scale=2.0)\n",
    "    kernel_regularizer = tf.contrib.layers.l2_regularizer(scale=params['l2_scale'])\n",
    "    training = (mode == tf.estimator.ModeKeys.TRAIN)\n",
    "    num_classes = 10\n",
    "\n",
    "    inputs = features['image'] * 2 - 1\n",
    "\n",
    "    hidden = tf.layers.conv2d(\n",
    "        inputs=inputs,\n",
    "        filters=16,\n",
    "        kernel_size=3,\n",
    "        kernel_initializer=kernel_initializer,\n",
    "        kernel_regularizer=kernel_regularizer,\n",
    "        padding='same')\n",
    "\n",
    "    hidden = resnet_layer(hidden, filters=16, strides=2, training=training, params=params)\n",
    "    hidden = resnet_layer(hidden, filters=32, strides=1, training=training, params=params)\n",
    "    hidden = resnet_layer(hidden, filters=64, strides=1, training=training, params=params)\n",
    "\n",
    "    hidden = batch_norm(\n",
    "        inputs=hidden,\n",
    "        momentum=params['bn_momentum'],\n",
    "        training=training)\n",
    "    hidden = tf.nn.relu(hidden)\n",
    "\n",
    "    hidden = tf.reduce_mean(hidden, axis=[1, 2])\n",
    "\n",
    "    logits = tf.layers.dense(\n",
    "        inputs=hidden,\n",
    "        units=num_classes,\n",
    "        kernel_initializer=kernel_initializer,\n",
    "        kernel_regularizer=kernel_regularizer)\n",
    "\n",
    "    outputs = tf.nn.softmax(logits)\n",
    "    prediction = tf.argmax(outputs, axis=-1)\n",
    "    \n",
    "    predictions = {\n",
    "        'prediction': prediction\n",
    "    }\n",
    "\n",
    "    loss = None\n",
    "    if mode != tf.estimator.ModeKeys.PREDICT:\n",
    "        loss = tf.losses.sparse_softmax_cross_entropy(\n",
    "            labels=labels['label'],\n",
    "            logits=logits)\n",
    "        loss = tf.losses.get_total_loss()\n",
    "        \n",
    "    train_op = None\n",
    "    if mode == tf.estimator.ModeKeys.TRAIN:\n",
    "        global_step = tf.train.get_or_create_global_step()\n",
    "        learning_rate = tf.train.piecewise_constant(\n",
    "            boundaries=params['lr_boundaries'],\n",
    "            values=params['lr_values'],\n",
    "            x=global_step)\n",
    "\n",
    "        optimizer = tf.train.MomentumOptimizer(\n",
    "            learning_rate=learning_rate,\n",
    "            momentum=params['momentum'])\n",
    "        with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):\n",
    "            train_op = optimizer.minimize(\n",
    "                global_step=global_step,\n",
    "                loss=loss)\n",
    "\n",
    "    eval_metric_ops = None\n",
    "    if mode != tf.estimator.ModeKeys.PREDICT:\n",
    "        eval_metric_ops = {\n",
    "            'accuracy': tf.metrics.accuracy(\n",
    "                predictions=predictions['prediction'],\n",
    "                labels=labels['label'])\n",
    "        }\n",
    "\n",
    "    estimator_spec = tf.estimator.EstimatorSpec(\n",
    "        loss=loss,\n",
    "        mode=mode,\n",
    "        train_op=train_op,\n",
    "        predictions=predictions,\n",
    "        eval_metric_ops=eval_metric_ops)\n",
    "    return estimator_spec"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Zx_enNtMKsuH"
   },
   "source": [
    "# Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 2434,
     "output_extras": [
      {},
      {},
      {}
     ]
    },
    "colab_type": "code",
    "id": "nvBISqbuDu1k",
    "outputId": "9401f25b-0cc5-4f29-c2a0-ebcfa2705a92"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Directory: jobs/1522769570\n",
      "Hyperparameters: {'lr_boundaries': [1000, 2000], 'momentum': 0.9, 'bn_momentum': 0.99, 'l2_scale': 0.0001, 'lr_values': [0.1, 0.01, 0.001]}\n",
      "INFO:tensorflow:Using config: {'_save_checkpoints_secs': None, '_session_config': None, '_keep_checkpoint_max': 5, '_task_type': 'worker', '_global_id_in_cluster': 0, '_is_chief': True, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x7f1e483364d0>, '_evaluation_master': '', '_save_checkpoints_steps': 1000, '_keep_checkpoint_every_n_hours': 10000, '_service': None, '_num_ps_replicas': 0, '_tf_random_seed': None, '_master': '', '_num_worker_replicas': 1, '_task_id': 0, '_log_step_count_steps': 100, '_model_dir': 'jobs/1522769570', '_save_summary_steps': 100}\n",
      "INFO:tensorflow:Running training and evaluation locally (non-distributed).\n",
      "INFO:tensorflow:Start train and evaluate loop. The evaluate will happen after 300 secs (eval_spec.throttle_secs) or training is finished.\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Saving checkpoints for 1 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:loss = 2.5851061, step = 1\n",
      "INFO:tensorflow:global_step/sec: 2.41255\n",
      "INFO:tensorflow:loss = 2.0360875, step = 101 (41.451 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.50442\n",
      "INFO:tensorflow:loss = 1.9359419, step = 201 (39.940 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.60598\n",
      "INFO:tensorflow:loss = 1.7742838, step = 301 (38.364 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.5028\n",
      "INFO:tensorflow:loss = 1.7685987, step = 401 (39.957 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.61014\n",
      "INFO:tensorflow:loss = 1.7156684, step = 501 (38.313 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.50776\n",
      "INFO:tensorflow:loss = 1.5870397, step = 601 (39.879 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.6097\n",
      "INFO:tensorflow:loss = 1.4641153, step = 701 (38.317 sec)\n",
      "INFO:tensorflow:Saving checkpoints for 738 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 1.464026.\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Starting evaluation at 2018-04-03-15:37:57\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from jobs/1522769570/model.ckpt-738\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Evaluation [10/100]\n",
      "INFO:tensorflow:Evaluation [20/100]\n",
      "INFO:tensorflow:Evaluation [30/100]\n",
      "INFO:tensorflow:Evaluation [40/100]\n",
      "INFO:tensorflow:Finished evaluation at 2018-04-03-15:38:02\n",
      "INFO:tensorflow:Saving dict for global step 738: accuracy = 0.474, global_step = 738, loss = 1.7852933\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from jobs/1522769570/model.ckpt-738\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Saving checkpoints for 739 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.5266762, step = 739\n",
      "INFO:tensorflow:global_step/sec: 2.43879\n",
      "INFO:tensorflow:loss = 1.5383127, step = 839 (41.010 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.50739\n",
      "INFO:tensorflow:loss = 1.3913659, step = 939 (39.884 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.60719\n",
      "INFO:tensorflow:loss = 1.1947991, step = 1039 (38.351 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.50678\n",
      "INFO:tensorflow:loss = 1.2099195, step = 1139 (39.895 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.61037\n",
      "INFO:tensorflow:loss = 1.0639611, step = 1239 (38.310 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.50781\n",
      "INFO:tensorflow:loss = 1.1801223, step = 1339 (39.872 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.6106\n",
      "INFO:tensorflow:loss = 1.0911791, step = 1439 (38.306 sec)\n",
      "INFO:tensorflow:Saving checkpoints for 1479 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 1.1021699.\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Starting evaluation at 2018-04-03-15:43:09\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from jobs/1522769570/model.ckpt-1479\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Evaluation [10/100]\n",
      "INFO:tensorflow:Evaluation [20/100]\n",
      "INFO:tensorflow:Evaluation [30/100]\n",
      "INFO:tensorflow:Evaluation [40/100]\n",
      "INFO:tensorflow:Finished evaluation at 2018-04-03-15:43:13\n",
      "INFO:tensorflow:Saving dict for global step 1479: accuracy = 0.641, global_step = 1479, loss = 1.168431\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from jobs/1522769570/model.ckpt-1479\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Saving checkpoints for 1480 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.1032535, step = 1480\n",
      "INFO:tensorflow:global_step/sec: 2.40106\n",
      "INFO:tensorflow:loss = 1.0231119, step = 1580 (41.650 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.50714\n",
      "INFO:tensorflow:loss = 1.1431658, step = 1680 (39.892 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.61301\n",
      "INFO:tensorflow:loss = 1.0614495, step = 1780 (38.268 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.5223\n",
      "INFO:tensorflow:loss = 1.0840397, step = 1880 (39.649 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.62217\n",
      "INFO:tensorflow:loss = 1.0816431, step = 1980 (38.132 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.51904\n",
      "INFO:tensorflow:loss = 0.9814845, step = 2080 (39.697 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.61773\n",
      "INFO:tensorflow:loss = 0.9994856, step = 2180 (38.205 sec)\n",
      "INFO:tensorflow:Saving checkpoints for 2220 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 0.9882607.\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Starting evaluation at 2018-04-03-15:48:19\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from jobs/1522769570/model.ckpt-2220\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Evaluation [10/100]\n",
      "INFO:tensorflow:Evaluation [20/100]\n",
      "INFO:tensorflow:Evaluation [30/100]\n",
      "INFO:tensorflow:Evaluation [40/100]\n",
      "INFO:tensorflow:Finished evaluation at 2018-04-03-15:48:24\n",
      "INFO:tensorflow:Saving dict for global step 2220: accuracy = 0.6912, global_step = 2220, loss = 1.0292851\n",
      "INFO:tensorflow:Calling model_fn.\n",
      "INFO:tensorflow:Done calling model_fn.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Graph was finalized.\n",
      "INFO:tensorflow:Restoring parameters from jobs/1522769570/model.ckpt-2220\n",
      "INFO:tensorflow:Running local_init_op.\n",
      "INFO:tensorflow:Done running local_init_op.\n",
      "INFO:tensorflow:Saving checkpoints for 2221 into jobs/1522769570/model.ckpt.\n",
      "INFO:tensorflow:loss = 0.98717695, step = 2221\n",
      "INFO:tensorflow:global_step/sec: 2.4809\n",
      "INFO:tensorflow:loss = 1.0000134, step = 2321 (40.313 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.51872\n",
      "INFO:tensorflow:loss = 0.9771461, step = 2421 (39.707 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.62311\n",
      "INFO:tensorflow:loss = 0.981297, step = 2521 (38.118 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.51619\n",
      "INFO:tensorflow:loss = 0.9755035, step = 2621 (39.740 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.62201\n",
      "INFO:tensorflow:loss = 0.926662, step = 2721 (38.141 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.5176\n",
      "INFO:tensorflow:loss = 0.9588964, step = 2821 (39.723 sec)\n",
      "INFO:tensorflow:global_step/sec: 2.62522\n",
      "INFO:tensorflow:loss = 0.99170995, step = 2921 (38.091 sec)\n"
     ]
    }
   ],
   "source": [
    "max_steps_train = 3000\n",
    "batch_size = 256\n",
    "model_dir = os.path.join('jobs/', str(int(time.time())))\n",
    "print('Model Directory:', model_dir)\n",
    "\n",
    "params = {\n",
    "    'l2_scale': 1e-4,\n",
    "    'lr_values': [1e-1, 1e-2, 1e-3],\n",
    "    'lr_boundaries': [max_steps_train // 3, max_steps_train // 3 * 2],\n",
    "    'bn_momentum': 0.99,\n",
    "    'momentum': 0.9,\n",
    "}\n",
    "print('Hyperparameters:', params)\n",
    "\n",
    "config = tf.estimator.RunConfig(\n",
    "    save_checkpoints_steps=1000,\n",
    "    save_checkpoints_secs=None,\n",
    "    save_summary_steps=100)\n",
    "\n",
    "tf.logging.set_verbosity(tf.logging.INFO)\n",
    "\n",
    "estimator = tf.estimator.Estimator(\n",
    "    model_fn=model_fn,\n",
    "    model_dir=model_dir,\n",
    "    config=config,\n",
    "    params=params)\n",
    "\n",
    "filenames_train = [\n",
    "    os.path.join(data_dir, 'data_batch_{}.bin'.format(i))\n",
    "    for i in xrange(1, 6)]\n",
    "filenames_eval = [os.path.join(data_dir, 'test_batch.bin')]\n",
    "\n",
    "input_fn_train = get_input_fn(\n",
    "    filenames=filenames_train,\n",
    "    batch_size=batch_size,\n",
    "    num_epochs=None,\n",
    "    shuffle=True,\n",
    "    augment=True)\n",
    "input_fn_eval = get_input_fn(\n",
    "    filenames=filenames_eval,\n",
    "    batch_size=batch_size,\n",
    "    num_epochs=1,\n",
    "    shuffle=False,\n",
    "    augment=False)\n",
    "\n",
    "train_spec = tf.estimator.TrainSpec(\n",
    "    input_fn=input_fn_train,\n",
    "    max_steps=max_steps_train)\n",
    "eval_spec = tf.estimator.EvalSpec(\n",
    "    input_fn=input_fn_eval,\n",
    "    start_delay_secs=0,\n",
    "    throttle_secs=60 * 5)\n",
    "\n",
    "tf.estimator.train_and_evaluate(\n",
    "    train_spec=train_spec,\n",
    "    eval_spec=eval_spec,\n",
    "    estimator=estimator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "otL8RWiLDy4l"
   },
   "source": [
    "# Download\n",
    "\n",
    "You can download the model directory for inspection via TensorBoard."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 34,
     "output_extras": [
      {}
     ]
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 12500,
     "status": "ok",
     "timestamp": 1522768166850,
     "user": {
      "displayName": "Jim Fleming",
      "photoUrl": "//lh5.googleusercontent.com/-bMUyWrg0nTk/AAAAAAAAAAI/AAAAAAAAADA/bxg68Muit30/s50-c-k-no/photo.jpg",
      "userId": "110726001486577440958"
     },
     "user_tz": 420
    },
    "id": "7qjPqmKV_Qtr",
    "outputId": "0ec5b777-4ca5-422b-8304-f91f3e9a2f16"
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'model_dir' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-25e5930d6818>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmodel_tarball\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel_dir\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m'.tar.gz'\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;32mwith\u001b[0m \u001b[0mtarfile\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_tarball\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"w:gz\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mtar\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mtar\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_dir\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marcname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mos\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbasename\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_dir\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mfiles\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdownload\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_tarball\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'model_dir' is not defined"
     ]
    }
   ],
   "source": [
    "model_tarball = model_dir + '.tar.gz'\n",
    "with tarfile.open(model_tarball, \"w:gz\") as tar:\n",
    "    tar.add(model_dir, arcname=os.path.basename(model_dir))\n",
    "files.download(model_tarball)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "ENj54x_6DdlS"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "default_view": {},
   "name": "Nitish-Gaddam-CIFAR-10.ipynb",
   "provenance": [
    {
     "file_id": "1jYL_Vi3OtTHcjj8HupGgLz6r-IjepS8P",
     "timestamp": 1522775566920
    }
   ],
   "toc_visible": true,
   "version": "0.3.2",
   "views": {}
  },
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}