Food image classification with CNN in Tensorflow.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "from scipy.misc import imread, imresize, imsave\n",
    "from random import shuffle\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "#import glop\n",
    "#import h5py\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATASET_DIR = 'Gaze_UPMC_Food20/images'\n",
    "TRAIN_DIR = 'dataset/train_set/'\n",
    "TEST_DIR = 'dataset/test_set/'\n",
    "\n",
    "IMAGE_SIZE = 128\n",
    "LR = 1e-4\n",
    "CLASSES = 5\n",
    "class_name = ['apple-pie', 'chocolate-cake', 'donuts', 'french-toast', 'gnocchi']\n",
    "\n",
    "\n",
    "MODEL_NAME = 'Food-image-classifer-{}-{}-model'.format(LR, '2conv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lable the image\n",
    "def label_img(index):\n",
    "    label = [0 for _ in range(CLASSES)]\n",
    "    label[index] = 1\n",
    "    return label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/tenzin/ML/project/food-image-classifier/.env/lib/python3.5/site-packages/ipykernel_launcher.py:8: DeprecationWarning: `imread` is deprecated!\n",
      "`imread` is deprecated in SciPy 1.0.0, and will be removed in 1.2.0.\n",
      "Use ``imageio.imread`` instead.\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "# label train and test dataset and save them in numpy array\n",
    "def create_dataset(path, dataset_name):\n",
    "    dataset = []\n",
    "    for index, img_dir_name in enumerate(sorted(os.listdir(path)[:5])):\n",
    "        dir_path = os.path.join(path, img_dir_name)\n",
    "        for img in os.listdir(dir_path):\n",
    "            img_path = os.path.join(dir_path, img)\n",
    "            img = imread(img_path)\n",
    "            dataset.append((img, np.array(label_img(index))))\n",
    "        shuffle(dataset)\n",
    "        np.save('dataset/' + dataset_name + '.npy', dataset)\n",
    "\n",
    "# Label train set\n",
    "create_dataset(TRAIN_DIR, 'train_set')\n",
    "# Label test set\n",
    "create_dataset(TEST_DIR, 'test_set')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## List images and label them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "shuffle_data = True  # shuffle the addresses before saving\n",
    "hdf5_path = '/dataset/dataset.hdf5'  # address to where you want to save the hdf5 file\n",
    "train_path = 'dataset/train_set/'\n",
    "test_pat = 'dataset/test_set'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dataset Description\n",
    "\n",
    "- dataset[]\n",
    "   - num_images = 800, 10 class 0f each 80 images\n",
    "   - 0 - img and 1 - label\n",
    "   - img: \n",
    "       - [x, y, z] \n",
    "   - label:\n",
    "       - [0, 0, 0, 0, 0, 1, 0, 0, 0, 0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Helper function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_train():\n",
    "    return np.load('dataset/train_set.npy')\n",
    "\n",
    "def load_test():\n",
    "    return np.load('dataset/test_set.npy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training Datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_set = load_train()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(400, 2)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_set.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f07b5e459e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(train_set[0][0][:, :, 0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X = np.array([i[0] for i in train_set]).reshape(-1, IMAGE_SIZE, IMAGE_SIZE, 3)\n",
    "train_y = np.array([list(i[1]) for i in train_set]) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train_X shape:  (400, 128, 128, 3)\n",
      "Train_y shape:  (400, 5)\n"
     ]
    }
   ],
   "source": [
    "print(\"Train_X shape: \", train_X.shape)\n",
    "print(\"Train_y shape: \", train_y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load train batch\n",
    "def train_batch(size=16):\n",
    "    first_index,last_index = 0, size\n",
    "    while last_index <= len(train_X):\n",
    "        # print('First Index: ', first_index)\n",
    "        # print('Last Index: ', last_index)\n",
    "        yield train_X[first_index: last_index], train_y[first_index: last_index]\n",
    "        first_index, last_index = last_index, last_index + size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f07b1758978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(train_X[1][:, :, 0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'donuts'"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_name[np.argmax(train_y[1])]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing Datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_set = load_test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 2)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_set.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_X = np.array([i[0] for i in test_set]).reshape(-1, IMAGE_SIZE, IMAGE_SIZE, 3)\n",
    "test_y = np.array([list(i[1]) for i in test_set]) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test_X shape:  (100, 128, 128, 3)\n",
      "Test_y shape:  (100, 5)\n"
     ]
    }
   ],
   "source": [
    "print(\"Test_X shape: \", test_X.shape)\n",
    "print(\"Test_y shape: \", test_y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f07b3cc2b70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(test_X[0][:, :, 0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'french-toast'"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_name[np.argmax(test_y[0])]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Place holder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "tf.reset_default_graph()\n",
    "X = tf.placeholder(tf.float32, shape=[None, IMAGE_SIZE, IMAGE_SIZE, 3], name='X')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_true = tf.placeholder(tf.float32, shape=[None, CLASSES], name=\"y_true\")\n",
    "y_true_cls = tf.argmax(y_true, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## CNN Model - Google Deep MNIST Architectuer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Weight Initialization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "def weight_variable(name, shape):\n",
    "    W = tf.get_variable(name, shape,\n",
    "           initializer=tf.contrib.layers.xavier_initializer())\n",
    "    return W\n",
    "\n",
    "def bias_variable(shape):\n",
    "    initial = tf.constant(0.5, shape=shape)\n",
    "    return tf.Variable(initial)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Convolution and Pooling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "def conv2d(x, W):\n",
    "    return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding=\"SAME\")\n",
    "\n",
    "def max_pool_2x2(x):\n",
    "    return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], \n",
    "                          strides=[1, 2, 2, 1], padding=\"SAME\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "h_pool3 shape:  (?, 16, 16, 128)\n",
      "h_fc1 shape:  (?, 1024)\n",
      "y_conv shape:  (?, 5)\n"
     ]
    }
   ],
   "source": [
    "# Frist Conv layer\n",
    "W_conv1 = weight_variable('W_conv1', [5, 5, 3, 32])\n",
    "b_conv1 = bias_variable([32])\n",
    "\n",
    "h_conv1 = tf.nn.relu(conv2d(X, W_conv1) + b_conv1)\n",
    "h_pool1 = max_pool_2x2(h_conv1)\n",
    "\n",
    "# Second Conv Layer\n",
    "W_conv2 = weight_variable('W_conv2', [5, 5, 32, 64])\n",
    "b_conv2 = bias_variable([64])\n",
    "\n",
    "h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)\n",
    "h_pool2 = max_pool_2x2(h_conv2)\n",
    "\n",
    "# Third Conv Layer\n",
    "W_conv3 = weight_variable('W_conv3', [5, 5, 64, 128])\n",
    "b_conv3 = bias_variable([128])\n",
    "\n",
    "h_conv3 = tf.nn.relu(conv2d(h_pool2, W_conv3) + b_conv3)\n",
    "h_pool3 = max_pool_2x2(h_conv3)\n",
    "\n",
    "print(\"h_pool3 shape: \", h_pool3.get_shape())\n",
    "\n",
    "# Fully or Densely Connected Layer\n",
    "W_fc1 = weight_variable('W_fc1', [16*16*128, 1024])\n",
    "b_fc1 = bias_variable([1024])\n",
    "\n",
    "h_pool3_flat = tf.reshape(h_pool3, [-1, 16*16*128])\n",
    "h_fc1 = tf.nn.relu(tf.matmul(h_pool3_flat, W_fc1) + b_fc1)\n",
    "\n",
    "print(\"h_fc1 shape: \", h_fc1.get_shape())\n",
    "\n",
    "# Dropout\n",
    "keep_prob = tf.placeholder(tf.float32)\n",
    "h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)\n",
    "\n",
    "# Readout Layer\n",
    "W_fc2 = weight_variable('W_fc2', [1024, 5])\n",
    "b_fc2 = bias_variable([5])\n",
    "\n",
    "y_conv = tf.add(tf.matmul(h_fc1_drop, W_fc2), b_fc2, name='y_conv')\n",
    "\n",
    "print(\"y_conv shape: \", y_conv.get_shape())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feed forward - Prediction\n",
    "y_pred = tf.nn.softmax(y_conv, name='y_pred')\n",
    "y_pred_cls = tf.argmax(y_pred, axis=1, name='y_pred_cls')\n",
    "\n",
    "# Training Objective\n",
    "cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=y_conv, labels=y_true)\n",
    "#cost = tf.reduce_mean(cross_entropy)\n",
    "train_step = tf.train.AdamOptimizer(LR).minimize(cross_entropy)\n",
    "\n",
    "# Evaluation\n",
    "correct_prediction = tf.equal(y_pred_cls, y_true_cls)\n",
    "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "saver = tf.train.Saver()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "sess = tf.Session()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 0, training accuracy 0.25\n",
      "step 100, training accuracy 1\n",
      "step 200, training accuracy 1\n",
      "step 300, training accuracy 1\n",
      "step 400, training accuracy 1\n",
      "step 500, training accuracy 1\n",
      "step 600, training accuracy 1\n",
      "step 700, training accuracy 1\n",
      "step 800, training accuracy 1\n",
      "step 900, training accuracy 1\n",
      "Test Accuracy: 0.45\n"
     ]
    }
   ],
   "source": [
    "counter = 0\n",
    "sess.run(tf.global_variables_initializer())\n",
    "for i in range(1000):\n",
    "    for batch in train_batch():\n",
    "        train_step.run(session=sess, feed_dict={X: batch[0], y_true: batch[1], keep_prob: 0.5})\n",
    "    if i % 100 == 0:\n",
    "        train_accuracy = accuracy.eval(session=sess, feed_dict={\n",
    "                X: batch[0], y_true: batch[1], keep_prob: 1.0})\n",
    "        print('step %d, training accuracy %g' % (i, train_accuracy))\n",
    "        counter += 1\n",
    "test_accuracy = accuracy.eval(session=sess, feed_dict={X: test_X, y_true: test_y, keep_prob: 1.0})\n",
    "print(\"Test Accuracy: %g\" % test_accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'model/Food-image-classifer-0.0001-2conv-model'"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "saver.save(sess, 'model/' + MODEL_NAME)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAD8CAYAAACcjGjIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAGTdJREFUeJzt3X+QXWWd5/H3p2/f7tsLiQRoY8yPScTgEi0nai+yNTrF\nyACBtQxuTbnJzgwZlzJaQpXWztZsmP0D11mqnN1Rd6lhmYqSItQqkRUZUlNxMZOlZKZq0HQkCwk/\nNk2AJT0hiYkKCunQyXf/uM/tnPS9t7vt292308/nVXWrz/2eH/c5seXTz/Occ48iAjMzs6KOdjfA\nzMxmH4eDmZnVcTiYmVkdh4OZmdVxOJiZWR2Hg5mZ1XE4mJlZHYeDmZnVcTiYmVmdznY3YLIuvfTS\nWL58ebubYWZ2XtmzZ89PI6J3vO3O23BYvnw5/f397W6Gmdl5RdLLE9nOw0pmZlbH4WBmZnUcDmZm\nVsfhYGZmdRwOZmZWx+FgZmZ1HA5mZlYnu3B4ZO8g/+OJCV3ma2aWrezC4ftPv8q9f/9iu5thZjar\nZRcOVyyaz0vHf8Ubp4bb3RQzs1lr3HCQtEXSUUn7CrXvSNqbXi9J2pvqyyW9WVj3V4V9PiTpaUkD\nku6SpFS/WNJOSQfSzwXTcaI1VyyaRwQ89+rr0/kxZmbntYn0HO4D1hQLEfGvImJ1RKwGHgK+V1j9\nQm1dRHyuUL8H+AywMr1qx9wE7IqIlcCu9H7aXLFoPgDPHn5tOj/GzOy8Nm44RMTjwIlG69Jf/58C\nHhjrGJIWAfMj4omICOB+4Ka0ei2wNS1vLdSnxZIFPcyrdPLMPzoczMyaaXXO4aPAkYg4UKitkPSk\npB9K+miqLQYOFbY5lGoACyPicFp+FVjYYpvGJIkr3jHfPQczszG0Gg7rObfXcBhYFhEfAP4t8G1J\n8yd6sNSriGbrJW2U1C+p/9ixY5NtM1csmsdzr77OmTNNP8rMLGuTDgdJncC/BL5Tq0XEUEQcT8t7\ngBeAy4FBYElh9yWpBnAkDTvVhp+ONvvMiNgcEX0R0dfbO+6zKpq6YtF83jh1mv934o1JH8PMbC5r\npefwu8BzETEyXCSpV1IpLb+L6sTzwTRs9Jqkq9I8xc3AI2m37cCGtLyhUJ82q97pSWkzs7FM5FLW\nB4B/AN4j6ZCkW9KqddRPRP828FS6tPW7wOciojaZ/Xngm8AA1R7F91P9K8C1kg5QDZyvtHA+E3L5\nwnl0yOFgZtbMuI8JjYj1Tep/1KD2ENVLWxtt3w+8r0H9OHDNeO2YSpVyiXf1Xsgzh32vg5lZI9nd\nIV1zxSJfsWRm1kzG4TCPwZ+/yS/efKvdTTEzm3UyDgdPSpuZNZNtOKxyOJiZNZVtOLx9XjeXXNDl\ncDAzayDbcJCUJqV9xZKZ2WjZhgNUJ6WfP/I6w6fPtLspZmazSubhMJ9Tw2d48ae/andTzMxmlezD\nAeAZzzuYmZ0j63C4rPdCyiV53sHMbJSsw6Grs4N3v32er1gyMxsl63CA6qS0h5XMzM6VfTisWjSf\nY68P8dNfDrW7KWZms4bDwXdKm5nVyT4c/B1LZmb1sg+HBRd08Y75FV+xZGZWkH04QHVS2j0HM7Oz\nHA5Uh5YGjv6SoeHT7W6Kmdms4HCgGg7DZ4KBo79sd1PMzGaFccNB0hZJRyXtK9S+JGlQ0t70urGw\n7nZJA5Kel3R9ob4m1QYkbSrUV0j6Uap/R1LXVJ7gRIx8jcY/emjJzAwm1nO4D1jToP71iFidXjsA\nJK0C1gHvTfv8d0klSSXgbuAGYBWwPm0L8OfpWO8Gfgbc0soJTcaKSy+gUu7wpLSZWTJuOETE48CJ\nCR5vLbAtIoYi4kVgALgyvQYi4mBEnAK2AWslCfgY8N20/1bgpl/zHFpW6hDvecd8T0qbmSWtzDnc\nJumpNOy0INUWA68UtjmUas3qlwA/j4jhUfWGJG2U1C+p/9ixYy00vd5lvRfw8nF/dbeZGUw+HO4B\nLgNWA4eBr05Zi8YQEZsjoi8i+np7e6f02Bd2d/LmW75aycwMoHMyO0XEkdqypG8Af5PeDgJLC5su\nSTWa1I8DF0nqTL2H4vYzqlIuORzMzJJJ9RwkLSq8/SRQu5JpO7BOUrekFcBK4MfAbmBlujKpi+qk\n9faICOAx4PfS/huARybTplZVyiVOvnWGapPMzPI2bs9B0gPA1cClkg4BdwBXS1oNBPAS8FmAiNgv\n6UHgGWAYuDUiTqfj3AY8CpSALRGxP33Evwe2SfpPwJPAvVN2dr+GSrmak0PDZ6iUS+1ogpnZrDFu\nOETE+gblpv8Bj4g7gTsb1HcAOxrUD1K9mqmtelIgnHzrtMPBzLLnO6STWiB43sHMzOEwojasdPKt\nM21uiZlZ+zkcktqw0pun3HMwM3M4JN21OQd/M6uZmcOhpjghbWaWO4dDUnE4mJmNcDgkZ3sOnpA2\nM3M4JLWrlTwhbWbmcBjR4wlpM7MRDoek25eympmNcDgktZ7D0LDnHMzMHA5JuSQ65J6DmRk4HEZI\noqdc8qWsZmY4HM7hB/6YmVU5HApqD/wxM8udw6GgUu7wsJKZGQ6Hc/R0ec7BzAwcDueodHrOwcwM\nJhAOkrZIOippX6H2XyQ9J+kpSQ9LuijVl0t6U9Le9Pqrwj4fkvS0pAFJd0lSql8saaekA+nnguk4\n0Ymo+GolMzNgYj2H+4A1o2o7gfdFxPuB/wvcXlj3QkSsTq/PFer3AJ8BVqZX7ZibgF0RsRLYld63\nRfVqJU9Im5mNGw4R8ThwYlTtBxExnN4+ASwZ6xiSFgHzI+KJiAjgfuCmtHotsDUtby3UZ1yl3MGQ\new5mZlMy5/BvgO8X3q+Q9KSkH0r6aKotBg4VtjmUagALI+JwWn4VWDgFbZoU3wRnZlbV2crOkv4D\nMAx8K5UOA8si4rikDwF/Lem9Ez1eRISkGOPzNgIbAZYtWzb5hjfhm+DMzKom3XOQ9EfAx4HfT0NF\nRMRQRBxPy3uAF4DLgUHOHXpakmoAR9KwU2346Wizz4yIzRHRFxF9vb29k216U9VLWT3nYGY2qXCQ\ntAb4E+ATEfFGod4rqZSW30V14vlgGjZ6TdJV6Sqlm4FH0m7bgQ1peUOhPuMqnR28+dZpUtaZmWVr\nIpeyPgD8A/AeSYck3QL8JTAP2DnqktXfBp6StBf4LvC5iKhNZn8e+CYwQLVHUZun+ApwraQDwO+m\n921R6fLXdpuZwQTmHCJifYPyvU22fQh4qMm6fuB9DerHgWvGa8dMqHTWniN9mkp6voOZWY58h3RB\nT1ctHNxzMLO8ORwKKuXqP4evWDKz3DkcCmqPCvW9DmaWO4dDQXcKB/cczCx3DocC9xzMzKocDgUV\nh4OZGeBwOEdtQtpXK5lZ7hwOBbVhpTdPuedgZnlzOBSMDCsNOxzMLG8Oh4KKew5mZoDD4Ry1OQd/\nt5KZ5c7hUNBV6qBDvlrJzMzhUCCp+sAfDyuZWeYcDqP0lEuekDaz7DkcRqn2HDznYGZ5cziMUil3\nuOdgZtlzOIxSKZc46TkHM8ucw2EUzzmYmU0wHCRtkXRU0r5C7WJJOyUdSD8XpLok3SVpQNJTkj5Y\n2GdD2v6ApA2F+ockPZ32uUuSpvIkfx2+WsnMbOI9h/uANaNqm4BdEbES2JXeA9wArEyvjcA9UA0T\n4A7gw8CVwB21QEnbfKaw3+jPmjGVcslfvGdm2ZtQOETE48CJUeW1wNa0vBW4qVC/P6qeAC6StAi4\nHtgZESci4mfATmBNWjc/Ip6IiADuLxxrxlXKHb4Jzsyy18qcw8KIOJyWXwUWpuXFwCuF7Q6l2lj1\nQw3qbVHtOTgczCxvUzIhnf7ij6k41lgkbZTUL6n/2LFj0/IZPeWSHxNqZtlrJRyOpCEh0s+jqT4I\nLC1styTVxqovaVCvExGbI6IvIvp6e3tbaHpz1WElzzmYWd5aCYftQO2Kow3AI4X6zemqpauAX6Th\np0eB6yQtSBPR1wGPpnWvSboqXaV0c+FYM67Wc6h2hszM8tQ5kY0kPQBcDVwq6RDVq46+Ajwo6Rbg\nZeBTafMdwI3AAPAG8GmAiDgh6c+A3Wm7L0dEbZL781SviOoBvp9ebdGdnukwNHxm5PkOZma5mVA4\nRMT6JquuabBtALc2Oc4WYEuDej/wvom0ZbrVHhV68q3TDgczy5bvkB5l5FGhnncws4w5HEbp6ar+\nk/hyVjPLmcNhlEpneo60w8HMMuZwGKXSdXbOwcwsVw6HUdxzMDNzONTpST2HIU9Im1nGHA6jVMrV\nfxL3HMwsZw6HUYr3OZiZ5crhMErtPgf3HMwsZw6HUXwTnJmZw6FObc7Bw0pmljOHwyhdpQ4kh4OZ\n5c3hMIqk6td2n3I4mFm+HA4NVMolTg47HMwsXw6HBqo9B09Im1m+HA4NdJc73HMws6w5HBroKZc4\n6TkHM8uYw6EBzzmYWe4cDg30lEu+Cc7MsjbpcJD0Hkl7C6/XJH1R0pckDRbqNxb2uV3SgKTnJV1f\nqK9JtQFJm1o9qVZVyh2+lNXMstY52R0j4nlgNYCkEjAIPAx8Gvh6RPxFcXtJq4B1wHuBdwJ/K+ny\ntPpu4FrgELBb0vaIeGaybWuVh5XMLHeTDodRrgFeiIiXJTXbZi2wLSKGgBclDQBXpnUDEXEQQNK2\ntG17w8E9BzPL2FTNOawDHii8v03SU5K2SFqQaouBVwrbHEq1ZvW26SmXODnsOQczy1fL4SCpC/gE\n8D9T6R7gMqpDToeBr7b6GYXP2iipX1L/sWPHpuqwdTznYGa5m4qeww3ATyLiCEBEHImI0xFxBvgG\nZ4eOBoGlhf2WpFqzep2I2BwRfRHR19vbOwVNb6wnzTlExLR9hpnZbDYV4bCewpCSpEWFdZ8E9qXl\n7cA6Sd2SVgArgR8Du4GVklakXsi6tG3bdJdLRMCQh5bMLFMtTUhLuoDqVUafLZT/s6TVQAAv1dZF\nxH5JD1KdaB4Gbo2I0+k4twGPAiVgS0Tsb6Vdrao98GforTMjy2ZmOWkpHCLiV8Alo2p/OMb2dwJ3\nNqjvAHa00pap1FN4VOjbKLe5NWZmM893SDfgp8GZWe4cDg0Uew5mZjlyODRQm2dwz8HMcuVwaKDi\nnoOZZc7h0EBtzmHI38xqZplyODTQ0+Weg5nlzeHQQKXTcw5mljeHQwO1noMf+GNmuXI4NFDrOXhY\nycxy5XBooNLlm+DMLG8Ohwa6Sh1IDgczy5fDoQFJ1a/tdjiYWaYcDk1UyiXPOZhZthwOTVR7Dr5a\nyczy5HBoorvc4Z6DmWXL4dBEpbPEkMPBzDLlcGiip8tzDmaWL4dDE5Vyh+cczCxbLYeDpJckPS1p\nr6T+VLtY0k5JB9LPBakuSXdJGpD0lKQPFo6zIW1/QNKGVtvVqp5yiTdPuedgZnmaqp7D70TE6ojo\nS+83AbsiYiWwK70HuAFYmV4bgXugGibAHcCHgSuBO2qB0i7d5RInhx0OZpan6RpWWgtsTctbgZsK\n9fuj6gngIkmLgOuBnRFxIiJ+BuwE1kxT2yakp1zipHsOZpapqQiHAH4gaY+kjam2MCIOp+VXgYVp\neTHwSmHfQ6nWrN42lXIHJ4c952BmeeqcgmN8JCIGJb0d2CnpueLKiAhJMQWfQwqfjQDLli2bikM2\n5TkHM8tZyz2HiBhMP48CD1OdMziShotIP4+mzQeBpYXdl6Ras/roz9ocEX0R0dfb29tq08dUSXMO\nEVOSa2Zm55WWwkHSBZLm1ZaB64B9wHagdsXRBuCRtLwduDldtXQV8Is0/PQocJ2kBWki+rpUa5tK\nuUQEnDrtoSUzy0+rw0oLgYcl1Y717Yj4X5J2Aw9KugV4GfhU2n4HcCMwALwBfBogIk5I+jNgd9ru\nyxFxosW2taRSTk+DO3WG7vTwHzOzXLQUDhFxEPjNBvXjwDUN6gHc2uRYW4AtrbRnKvXUwmH4NG+j\n3ObWmJnNLN8h3USlXP2n8aS0meXI4dBEsedgZpYbh0MTtTkH9xzMLEcOhya607CSv3zPzHLkcGhi\nZFjJX9ttZhlyODRRcTiYWcYcDk3Ueg5+4I+Z5cjh0MTZnoPnHMwsPw6HJtxzMLOcORyaOHu1ksPB\nzPLjcGiiu7MDyeFgZnlyODQhiUpnyeFgZllyOIyhp6vkOQczy5LDYQyVzg5frWRmWXI4jKHS5WEl\nM8uTw2EMnnMws1w5HMbQ01XysJKZZcnhMIZKucMT0maWJYfDGDysZGa5mnQ4SFoq6TFJz0jaL+kL\nqf4lSYOS9qbXjYV9bpc0IOl5SdcX6mtSbUDSptZOaepUfCmrmWWqs4V9h4E/joifSJoH7JG0M637\nekT8RXFjSauAdcB7gXcCfyvp8rT6buBa4BCwW9L2iHimhbZNiUpniSHPOZhZhiYdDhFxGDicll+X\n9CyweIxd1gLbImIIeFHSAHBlWjcQEQcBJG1L27Y9HHq6POdgZnmakjkHScuBDwA/SqXbJD0laYuk\nBam2GHilsNuhVGtWb/Q5GyX1S+o/duzYVDR9TJ5zMLNctRwOki4EHgK+GBGvAfcAlwGrqfYsvtrq\nZ9RExOaI6IuIvt7e3qk6bFO1r8+IiGn/LDOz2aSlcJBUphoM34qI7wFExJGIOB0RZ4BvcHboaBBY\nWth9Sao1q7ddpVwiAk6d9ryDmeWllauVBNwLPBsRXyvUFxU2+ySwLy1vB9ZJ6pa0AlgJ/BjYDayU\ntEJSF9VJ6+2TbddUGnka3CmHg5nlpZWrlX4L+EPgaUl7U+1PgfWSVgMBvAR8FiAi9kt6kOpE8zBw\na0ScBpB0G/AoUAK2RMT+Fto1ZSq1B/4Mn+ZtlNvcGjOzmdPK1Up/D6jBqh1j7HMncGeD+o6x9muX\nkUeFnvKktJnlxXdIj2FkWGnY4WBmeXE4jME9BzPLlcNhDN21OQffJW1mmXE4jKHHw0pmlimHwxjO\nXsrqcDCzvDgcxuCeg5nlyuEwhsrIhLTnHMwsLw6HMYzcBOcv3zOzzDgcxjDSc3A4mFlmHA5j6O7s\nQIIhh4OZZcbhMAZJVDr9qFAzy4/DYRyVcodvgjOz7DgcxtFTds/BzPLjcBhHpexHhZpZfhwO43A4\nmFmOHA7j8JyDmeXI4TCOni7POZhZfhwO46h0eljJzPIza8JB0hpJz0sakLSp3e2pqXQ5HMwsP7Mi\nHCSVgLuBG4BVwHpJq9rbqqpqz8FzDmaWl1kRDsCVwEBEHIyIU8A2YG2b2wRAT1eHew5mlp3Odjcg\nWQy8Unh/CPhwm9pyjkpniRNvnOLar/2w3U0xMwPg3g3/jGWX/JNp/YzZEg4TImkjsBFg2bJlM/KZ\nn1j9Tg6/dpKImJHPMzMbT1fn9A/6zJZwGASWFt4vSbVzRMRmYDNAX1/fjPzX+v1LLuLuf/3Bmfgo\nM7NZY7bMOewGVkpaIakLWAdsb3ObzMyyNSt6DhExLOk24FGgBGyJiP1tbpaZWbZmRTgARMQOYEe7\n22FmZrNnWMnMzGYRh4OZmdVxOJiZWR2Hg5mZ1XE4mJlZHZ2vd/5KOga8PMndLwV+OoXNOR/4nPPg\nc577Wj3f34iI3vE2Om/DoRWS+iOir93tmEk+5zz4nOe+mTpfDyuZmVkdh4OZmdXJNRw2t7sBbeBz\nzoPPee6bkfPNcs7BzMzGlmvPwczMxpBdOEhaI+l5SQOSNrW7PdNB0hZJRyXtK9QulrRT0oH0c0E7\n2ziVJC2V9JikZyTtl/SFVJ/L51yR9GNJ/yed839M9RWSfpR+v7+TvgJ/TpFUkvSkpL9J7+f0OUt6\nSdLTkvZK6k+1af/dziocJJWAu4EbgFXAekmr2tuqaXEfsGZUbROwKyJWArvS+7liGPjjiFgFXAXc\nmv53ncvnPAR8LCJ+E1gNrJF0FfDnwNcj4t3Az4Bb2tjG6fIF4NnC+xzO+XciYnXhEtZp/93OKhyA\nK4GBiDgYEaeAbcDaNrdpykXE48CJUeW1wNa0vBW4aUYbNY0i4nBE/CQtv071PxyLmdvnHBHxy/S2\nnF4BfAz4bqrPqXMGkLQE+BfAN9N7McfPuYlp/93OLRwWA68U3h9KtRwsjIjDaflVYGE7GzNdJC0H\nPgD8iDl+zml4ZS9wFNgJvAD8PCKG0yZz8ff7vwJ/ApxJ7y9h7p9zAD+QtEfSxlSb9t/tWfOwH5s5\nERGS5txlapIuBB4CvhgRr1X/qKyai+ccEaeB1ZIuAh4G/mmbmzStJH0cOBoReyRd3e72zKCPRMSg\npLcDOyU9V1w5Xb/bufUcBoGlhfdLUi0HRyQtAkg/j7a5PVNKUplqMHwrIr6XynP6nGsi4ufAY8A/\nBy6SVPujb679fv8W8AlJL1EdEv4Y8N+Y2+dMRAymn0ep/hFwJTPwu51bOOwGVqarG7qAdcD2Nrdp\npmwHNqTlDcAjbWzLlErjzvcCz0bE1wqr5vI596YeA5J6gGupzrU8Bvxe2mxOnXNE3B4RSyJiOdX/\n7/7viPh95vA5S7pA0rzaMnAdsI8Z+N3O7iY4STdSHbcsAVsi4s42N2nKSXoAuJrqtzceAe4A/hp4\nEFhG9dtsPxURoyetz0uSPgL8HfA0Z8ei/5TqvMNcPef3U52ILFH9I+/BiPiypHdR/av6YuBJ4A8i\nYqh9LZ0eaVjp30XEx+fyOadzezi97QS+HRF3SrqEaf7dzi4czMxsfLkNK5mZ2QQ4HMzMrI7DwczM\n6jgczMysjsPBzMzqOBzMzKyOw8HMzOo4HMzMrM7/B/h5fzuWLsGiAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f0c7525cac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(costs)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading model and using it for prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "float32\n",
      "(128, 128, 3)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel_launcher.py:3: DeprecationWarning: `imresize` is deprecated!\n",
      "`imresize` is deprecated in SciPy 1.0.0, and will be removed in 1.2.0.\n",
      "Use ``skimage.transform.resize`` instead.\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "img = Image.open('media_cdn/images/ngoenga.png')\n",
    "img = np.array(img)\n",
    "img = imresize(img, (128, 128)).astype(np.float32)\n",
    "\n",
    "print(img.dtype)\n",
    "print(img.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_dir = 'model/'\n",
    "model_meta_path = os.path.join(model_dir, MODEL_NAME + '.meta')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model/Food-image-classifer-0.0001-2conv-model\n"
     ]
    },
    {
     "ename": "InvalidArgumentError",
     "evalue": "You must feed a value for placeholder tensor 'Placeholder' with dtype float\n\t [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=<unknown>, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]\n\nCaused by op 'Placeholder', defined at:\n  File \"/usr/lib/python3.5/runpy.py\", line 184, in _run_module_as_main\n    \"__main__\", mod_spec)\n  File \"/usr/lib/python3.5/runpy.py\", line 85, in _run_code\n    exec(code, run_globals)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel_launcher.py\", line 16, in <module>\n    app.launch_new_instance()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/traitlets/config/application.py\", line 658, in launch_instance\n    app.start()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelapp.py\", line 477, in start\n    ioloop.IOLoop.instance().start()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/ioloop.py\", line 177, in start\n    super(ZMQIOLoop, self).start()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tornado/ioloop.py\", line 888, in start\n    handler_func(fd_obj, events)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tornado/stack_context.py\", line 277, in null_wrapper\n    return fn(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py\", line 440, in _handle_events\n    self._handle_recv()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py\", line 472, in _handle_recv\n    self._run_callback(callback, msg)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py\", line 414, in _run_callback\n    callback(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tornado/stack_context.py\", line 277, in null_wrapper\n    return fn(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelbase.py\", line 283, in dispatcher\n    return self.dispatch_shell(stream, msg)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelbase.py\", line 235, in dispatch_shell\n    handler(stream, idents, msg)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelbase.py\", line 399, in execute_request\n    user_expressions, allow_stdin)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/ipkernel.py\", line 196, in do_execute\n    res = shell.run_cell(code, store_history=store_history, silent=silent)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/zmqshell.py\", line 533, in run_cell\n    return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/IPython/core/interactiveshell.py\", line 2698, in run_cell\n    interactivity=interactivity, compiler=compiler, result=result)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/IPython/core/interactiveshell.py\", line 2802, in run_ast_nodes\n    if self.run_code(code, result):\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/IPython/core/interactiveshell.py\", line 2862, in run_code\n    exec(code_obj, self.user_global_ns, self.user_ns)\n  File \"<ipython-input-8-04376588bd65>\", line 2, in <module>\n    saver = tf.train.import_meta_graph(model_meta_path)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/training/saver.py\", line 1810, in import_meta_graph\n    **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/meta_graph.py\", line 660, in import_scoped_meta_graph\n    producer_op_list=producer_op_list)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/importer.py\", line 313, in import_graph_def\n    op_def=op_def)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/ops.py\", line 2956, in create_op\n    op_def=op_def)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/ops.py\", line 1470, in __init__\n    self._traceback = self._graph._extract_stack()  # pylint: disable=protected-access\n\nInvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'Placeholder' with dtype float\n\t [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=<unknown>, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]\n",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mInvalidArgumentError\u001b[0m                      Traceback (most recent call last)",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_do_call\u001b[0;34m(self, fn, *args)\u001b[0m\n\u001b[1;32m   1322\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1323\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1324\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0merrors\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mOpError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_run_fn\u001b[0;34m(session, feed_dict, fetch_list, target_list, options, run_metadata)\u001b[0m\n\u001b[1;32m   1301\u001b[0m                                    \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_list\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1302\u001b[0;31m                                    status, run_metadata)\n\u001b[0m\u001b[1;32m   1303\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/errors_impl.py\u001b[0m in \u001b[0;36m__exit__\u001b[0;34m(self, type_arg, value_arg, traceback_arg)\u001b[0m\n\u001b[1;32m    472\u001b[0m             \u001b[0mcompat\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_text\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc_api\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_Message\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstatus\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstatus\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 473\u001b[0;31m             c_api.TF_GetCode(self.status.status))\n\u001b[0m\u001b[1;32m    474\u001b[0m     \u001b[0;31m# Delete the underlying status object from memory otherwise it stays alive\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mInvalidArgumentError\u001b[0m: You must feed a value for placeholder tensor 'Placeholder' with dtype float\n\t [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=<unknown>, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mInvalidArgumentError\u001b[0m                      Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-20-8f410a77d634>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m     \u001b[0my_conv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_tensor_by_name\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'y_conv:0'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m     \u001b[0mpredict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msess\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_conv\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeed_dict\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mimg\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_conv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_shape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    887\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    888\u001b[0m       result = self._run(None, fetches, feed_dict, options_ptr,\n\u001b[0;32m--> 889\u001b[0;31m                          run_metadata_ptr)\n\u001b[0m\u001b[1;32m    890\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    891\u001b[0m         \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_run\u001b[0;34m(self, handle, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m   1118\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfinal_fetches\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mfinal_targets\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mhandle\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mfeed_dict_tensor\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1119\u001b[0m       results = self._do_run(handle, final_targets, final_fetches,\n\u001b[0;32m-> 1120\u001b[0;31m                              feed_dict_tensor, options, run_metadata)\n\u001b[0m\u001b[1;32m   1121\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1122\u001b[0m       \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_do_run\u001b[0;34m(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m   1315\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mhandle\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1316\u001b[0m       return self._do_call(_run_fn, self._session, feeds, fetches, targets,\n\u001b[0;32m-> 1317\u001b[0;31m                            options, run_metadata)\n\u001b[0m\u001b[1;32m   1318\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1319\u001b[0m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_do_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_prun_fn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhandle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeeds\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetches\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/ML/env/lib/python3.5/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_do_call\u001b[0;34m(self, fn, *args)\u001b[0m\n\u001b[1;32m   1334\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1335\u001b[0m           \u001b[0;32mpass\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1336\u001b[0;31m       \u001b[0;32mraise\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnode_def\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmessage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1337\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1338\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m_extend_graph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mInvalidArgumentError\u001b[0m: You must feed a value for placeholder tensor 'Placeholder' with dtype float\n\t [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=<unknown>, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]\n\nCaused by op 'Placeholder', defined at:\n  File \"/usr/lib/python3.5/runpy.py\", line 184, in _run_module_as_main\n    \"__main__\", mod_spec)\n  File \"/usr/lib/python3.5/runpy.py\", line 85, in _run_code\n    exec(code, run_globals)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel_launcher.py\", line 16, in <module>\n    app.launch_new_instance()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/traitlets/config/application.py\", line 658, in launch_instance\n    app.start()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelapp.py\", line 477, in start\n    ioloop.IOLoop.instance().start()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/ioloop.py\", line 177, in start\n    super(ZMQIOLoop, self).start()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tornado/ioloop.py\", line 888, in start\n    handler_func(fd_obj, events)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tornado/stack_context.py\", line 277, in null_wrapper\n    return fn(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py\", line 440, in _handle_events\n    self._handle_recv()\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py\", line 472, in _handle_recv\n    self._run_callback(callback, msg)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/zmq/eventloop/zmqstream.py\", line 414, in _run_callback\n    callback(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tornado/stack_context.py\", line 277, in null_wrapper\n    return fn(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelbase.py\", line 283, in dispatcher\n    return self.dispatch_shell(stream, msg)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelbase.py\", line 235, in dispatch_shell\n    handler(stream, idents, msg)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/kernelbase.py\", line 399, in execute_request\n    user_expressions, allow_stdin)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/ipkernel.py\", line 196, in do_execute\n    res = shell.run_cell(code, store_history=store_history, silent=silent)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/ipykernel/zmqshell.py\", line 533, in run_cell\n    return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/IPython/core/interactiveshell.py\", line 2698, in run_cell\n    interactivity=interactivity, compiler=compiler, result=result)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/IPython/core/interactiveshell.py\", line 2802, in run_ast_nodes\n    if self.run_code(code, result):\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/IPython/core/interactiveshell.py\", line 2862, in run_code\n    exec(code_obj, self.user_global_ns, self.user_ns)\n  File \"<ipython-input-8-04376588bd65>\", line 2, in <module>\n    saver = tf.train.import_meta_graph(model_meta_path)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/training/saver.py\", line 1810, in import_meta_graph\n    **kwargs)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/meta_graph.py\", line 660, in import_scoped_meta_graph\n    producer_op_list=producer_op_list)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/importer.py\", line 313, in import_graph_def\n    op_def=op_def)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/ops.py\", line 2956, in create_op\n    op_def=op_def)\n  File \"/home/tenzin/ML/env/lib/python3.5/site-packages/tensorflow/python/framework/ops.py\", line 1470, in __init__\n    self._traceback = self._graph._extract_stack()  # pylint: disable=protected-access\n\nInvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'Placeholder' with dtype float\n\t [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=<unknown>, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:                                                   \n",
    "    saver = tf.train.import_meta_graph(model_meta_path)                      \n",
    "    saver.restore(sess, tf.train.latest_checkpoint(model_dir))               \n",
    "    graph = tf.get_default_graph()                                           \n",
    "    X = graph.get_tensor_by_name('X:0')                                      \n",
    "    y_conv = graph.get_tensor_by_name('y_conv:0')                            \n",
    "    \n",
    "    predict = sess.run(y_conv, feed_dict={X: [img]}) \n",
    "    \n",
    "    print(y_conv.get_shape())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}