notebooks/vae-semi_supervised_learning/vae-m2-fit-mnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import time\n",
    "import h5py\n",
    "import keras\n",
    "import pandas as pd\n",
    "import math\n",
    "import joblib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from fuel.datasets.hdf5 import H5PYDataset\n",
    "\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.model_selection import StratifiedShuffleSplit\n",
    "from IPython.display import display\n",
    "\n",
    "from keras.layers import (Input, Dense, Lambda, Flatten, Reshape, BatchNormalization, Activation, \n",
    "                          Dropout, Conv2D, Conv2DTranspose, MaxPooling2D)\n",
    "from keras.layers.merge import concatenate\n",
    "from keras.regularizers import l2\n",
    "from keras.initializers import RandomUniform\n",
    "from keras.optimizers import RMSprop, Adam, SGD\n",
    "from keras.models import Model\n",
    "from keras import metrics\n",
    "from keras.utils import np_utils\n",
    "from keras import backend as K\n",
    "from keras_tqdm import TQDMNotebookCallback\n",
    "from keras.datasets import mnist"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Variational Autoencoder Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_rows, img_cols, img_chns = 28, 28, 1\n",
    "\n",
    "if K.image_data_format() == 'channels_first':\n",
    "    original_img_size = (img_chns, img_rows, img_cols)\n",
    "else:\n",
    "    original_img_size = (img_rows, img_cols, img_chns)\n",
    "\n",
    "batch_size = 200\n",
    "latent_dim = 128\n",
    "intermediate_dim = 512\n",
    "epsilon_std = 1.0\n",
    "epochs = 10\n",
    "activation = 'relu'\n",
    "dropout = 0.5\n",
    "learning_rate = 0.001\n",
    "decay = 0.0\n",
    "num_classes = 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load MNIST dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28, 1) (60000,)\n",
      "(10000, 28, 28, 1) (10000,)\n"
     ]
    }
   ],
   "source": [
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
    "X_train = X_train.reshape(X_train.shape[0], img_rows, img_cols, 1) / 255.\n",
    "X_test = X_test.reshape(X_test.shape[0], img_rows, img_cols, 1) / 255.\n",
    "\n",
    "print(X_train.shape, y_train.shape)\n",
    "print(X_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Encoder Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def create_enc_conv_layers(stage, **kwargs):\n",
    "    conv_name = '_'.join(['enc_conv', str(stage)])\n",
    "    bn_name = '_'.join(['enc_bn', str(stage)])\n",
    "    layers = [\n",
    "        Conv2D(name=conv_name, **kwargs),\n",
    "        BatchNormalization(name=bn_name),\n",
    "        Activation(activation),\n",
    "    ]\n",
    "    return layers\n",
    "\n",
    "def create_dense_layers(stage, width):\n",
    "    dense_name = '_'.join(['enc_conv', str(stage)])\n",
    "    bn_name = '_'.join(['enc_bn', str(stage)])\n",
    "    layers = [\n",
    "        Dense(width, name=dense_name),\n",
    "        BatchNormalization(name=bn_name),\n",
    "        Activation(activation),\n",
    "        Dropout(dropout),\n",
    "    ]\n",
    "    return layers\n",
    "\n",
    "def inst_layers(layers, in_layer):\n",
    "    x = in_layer\n",
    "    for layer in layers:\n",
    "        if isinstance(layer, list):\n",
    "            x = inst_layers(layer, x)\n",
    "        else:\n",
    "            x = layer(x)\n",
    "        \n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "enc_filters=64\n",
    "enc_layers = [\n",
    "    create_enc_conv_layers(stage=1, filters=enc_filters, kernel_size=3, strides=1, padding='same'),\n",
    "    create_enc_conv_layers(stage=2, filters=enc_filters, kernel_size=3, strides=1, padding='same'),\n",
    "    create_enc_conv_layers(stage=3, filters=enc_filters, kernel_size=3, strides=2, padding='same'),\n",
    "    Flatten(),\n",
    "    create_dense_layers(stage=4, width=intermediate_dim),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Labeled encoder\n",
    "x_in = Input(batch_shape=(batch_size,) + original_img_size)\n",
    "y_in = Input(batch_shape=(batch_size, num_classes))\n",
    "_enc_dense = inst_layers(enc_layers, x_in)\n",
    "\n",
    "_z_mean_1 = Dense(latent_dim)(_enc_dense)\n",
    "_z_log_var_1 = Dense(latent_dim)(_enc_dense)\n",
    "\n",
    "z_mean = _z_mean_1\n",
    "z_log_var = _z_log_var_1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reparameterization Trick"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sampling(args, batch_size=batch_size, latent_dim=latent_dim, epsilon_std=epsilon_std):\n",
    "    z_mean, z_log_var = args\n",
    "    \n",
    "    epsilon = K.random_normal(shape=(batch_size, latent_dim),\n",
    "                              mean=0., stddev=epsilon_std)\n",
    "    \n",
    "    return z_mean + K.exp(z_log_var) * epsilon\n",
    "\n",
    "z = Lambda(sampling, output_shape=(latent_dim,))([z_mean, z_log_var])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classifier Network "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier_layers = [\n",
    "    Conv2D(32, (3, 3), padding='same'),\n",
    "    Activation('relu'),\n",
    "    Conv2D(32, (3, 3)),\n",
    "    Activation('relu'),\n",
    "    MaxPooling2D(pool_size=(2, 2)),\n",
    "    Dropout(0.25),\n",
    "    Conv2D(64, (3, 3), padding='same'),\n",
    "    Activation('relu'),\n",
    "    Conv2D(64, (3, 3)),\n",
    "    Activation('relu'),\n",
    "    MaxPooling2D(pool_size=(2, 2)),\n",
    "    Dropout(0.25),\n",
    "    Flatten(),\n",
    "    Dense(512),\n",
    "    Activation('relu'),\n",
    "    Dropout(0.5),\n",
    "    Dense(num_classes),\n",
    "    Activation('softmax'),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "_cls_output = inst_layers(classifier_layers, x_in)\n",
    "_y_output = _cls_output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Decoder Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_dec_trans_conv_layers(stage, **kwargs):\n",
    "    conv_name = '_'.join(['dec_trans_conv', str(stage)])\n",
    "    bn_name = '_'.join(['dec_bn', str(stage)])\n",
    "    layers = [\n",
    "        Conv2DTranspose(name=conv_name, **kwargs),\n",
    "        BatchNormalization(name=bn_name),\n",
    "        Activation(activation),\n",
    "    ]\n",
    "    return layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "dec_filters = 64\n",
    "decoder_layers = [\n",
    "    create_dense_layers(stage=10, width=14 * 14 * 64),\n",
    "    Reshape((14, 14, 64)),\n",
    "    create_dec_trans_conv_layers(11, filters=dec_filters, kernel_size=3, strides=1, padding='same'),\n",
    "    create_dec_trans_conv_layers(12, filters=dec_filters, kernel_size=3, strides=1, padding='same'),\n",
    "    create_dec_trans_conv_layers(13, filters=dec_filters, kernel_size=3, strides=2, padding='same'),\n",
    "    Conv2DTranspose(name='x_decoded', filters=1, kernel_size=1, strides=1, activation='sigmoid'),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Labeled decoder\n",
    "_merged = concatenate([y_in, z])\n",
    "_dec_out = inst_layers(decoder_layers, _merged)\n",
    "_x_output = _dec_out\n",
    "\n",
    "# Unlabeled decoder\n",
    "u_merged = concatenate([_y_output, z])\n",
    "u_dec_out = inst_layers(decoder_layers, u_merged)\n",
    "u_x_output = u_dec_out"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loss Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "def kl_loss(x, x_decoded_mean, z_mean=z_mean, z_log_var=z_log_var):\n",
    "    kl_loss = - 0.5 * K.sum(1. + z_log_var - K.square(z_mean) - K.exp(z_log_var), axis=-1)\n",
    "   \n",
    "    return K.mean(kl_loss)\n",
    "\n",
    "def logxy_loss(x, x_decoded_mean):\n",
    "    x = K.flatten(x)\n",
    "    x_decoded_mean = K.flatten(x_decoded_mean)\n",
    "    xent_loss = img_rows * img_cols * img_chns * metrics.binary_crossentropy(x, x_decoded_mean)\n",
    "   \n",
    "    # p(y) for observed data is equally distributed\n",
    "    logy = np.log(1. / num_classes)\n",
    "    \n",
    "    return xent_loss - logy\n",
    "\n",
    "def labeled_vae_loss(x, x_decoded_mean):\n",
    "    return logxy_loss(x, x_decoded_mean) + kl_loss(x, x_decoded_mean)\n",
    "\n",
    "def cls_loss(y, y_pred, N=1000):\n",
    "    alpha = 0.1 * N\n",
    "    return alpha * metrics.categorical_crossentropy(y, y_pred)\n",
    "\n",
    "def unlabeled_vae_loss(x, x_decoded_mean):\n",
    "    entropy = metrics.categorical_crossentropy(_y_output, _y_output)\n",
    "    # This is probably not correct, see discussion here: https://github.com/bjlkeng/sandbox/issues/3\n",
    "    labeled_loss = logxy_loss(x, x_decoded_mean) + kl_loss(x, x_decoded_mean)\n",
    "    \n",
    "    return K.mean(K.sum(_y_output * labeled_loss, axis=-1)) + entropy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compile Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_1 (InputLayer)             (200, 28, 28, 1)      0                                            \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_1 (Conv2D)              (200, 28, 28, 64)     640         input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_1 (BatchNormalization)    (200, 28, 28, 64)     256         enc_conv_1[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "activation_1 (Activation)        (200, 28, 28, 64)     0           enc_bn_1[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_2 (Conv2D)              (200, 28, 28, 64)     36928       activation_1[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_2 (BatchNormalization)    (200, 28, 28, 64)     256         enc_conv_2[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "activation_2 (Activation)        (200, 28, 28, 64)     0           enc_bn_2[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_3 (Conv2D)              (200, 14, 14, 64)     36928       activation_2[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_3 (BatchNormalization)    (200, 14, 14, 64)     256         enc_conv_3[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "activation_3 (Activation)        (200, 14, 14, 64)     0           enc_bn_3[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "flatten_1 (Flatten)              (200, 12544)          0           activation_3[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_4 (Dense)               (200, 512)            6423040     flatten_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_4 (BatchNormalization)    (200, 512)            2048        enc_conv_4[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "activation_4 (Activation)        (200, 512)            0           enc_bn_4[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)              (200, 512)            0           activation_4[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dense_1 (Dense)                  (200, 128)            65664       dropout_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_2 (Dense)                  (200, 128)            65664       dropout_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)                (200, 28, 28, 32)     320         input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "input_2 (InputLayer)             (200, 10)             0                                            \n",
      "____________________________________________________________________________________________________\n",
      "lambda_1 (Lambda)                (200, 128)            0           dense_1[0][0]                    \n",
      "                                                                   dense_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "activation_5 (Activation)        (200, 28, 28, 32)     0           conv2d_1[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)      (200, 138)            0           input_2[0][0]                    \n",
      "                                                                   lambda_1[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)                (200, 26, 26, 32)     9248        activation_5[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_10 (Dense)              (200, 12544)          1743616     concatenate_1[0][0]              \n",
      "____________________________________________________________________________________________________\n",
      "activation_6 (Activation)        (200, 26, 26, 32)     0           conv2d_2[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_10 (BatchNormalization)   (200, 12544)          50176       enc_conv_10[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)   (200, 13, 13, 32)     0           activation_6[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "activation_11 (Activation)       (200, 12544)          0           enc_bn_10[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)              (200, 13, 13, 32)     0           max_pooling2d_1[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "dropout_5 (Dropout)              (200, 12544)          0           activation_11[0][0]              \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)                (200, 13, 13, 64)     18496       dropout_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "reshape_1 (Reshape)              (200, 14, 14, 64)     0           dropout_5[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "activation_7 (Activation)        (200, 13, 13, 64)     0           conv2d_3[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "dec_trans_conv_11 (Conv2DTranspo (200, 14, 14, 64)     36928       reshape_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)                (200, 11, 11, 64)     36928       activation_7[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dec_bn_11 (BatchNormalization)   (200, 14, 14, 64)     256         dec_trans_conv_11[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "activation_8 (Activation)        (200, 11, 11, 64)     0           conv2d_4[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "activation_12 (Activation)       (200, 14, 14, 64)     0           dec_bn_11[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2D)   (200, 5, 5, 64)       0           activation_8[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dec_trans_conv_12 (Conv2DTranspo (200, 14, 14, 64)     36928       activation_12[0][0]              \n",
      "____________________________________________________________________________________________________\n",
      "dropout_3 (Dropout)              (200, 5, 5, 64)       0           max_pooling2d_2[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "dec_bn_12 (BatchNormalization)   (200, 14, 14, 64)     256         dec_trans_conv_12[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "flatten_2 (Flatten)              (200, 1600)           0           dropout_3[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "activation_13 (Activation)       (200, 14, 14, 64)     0           dec_bn_12[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_3 (Dense)                  (200, 512)            819712      flatten_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dec_trans_conv_13 (Conv2DTranspo (200, 28, 28, 64)     36928       activation_13[0][0]              \n",
      "____________________________________________________________________________________________________\n",
      "activation_9 (Activation)        (200, 512)            0           dense_3[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "dec_bn_13 (BatchNormalization)   (200, 28, 28, 64)     256         dec_trans_conv_13[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_4 (Dropout)              (200, 512)            0           activation_9[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "activation_14 (Activation)       (200, 28, 28, 64)     0           dec_bn_13[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_4 (Dense)                  (200, 10)             5130        dropout_4[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "x_decoded (Conv2DTranspose)      (200, 28, 28, 1)      65          activation_14[0][0]              \n",
      "____________________________________________________________________________________________________\n",
      "activation_10 (Activation)       (200, 10)             0           dense_4[0][0]                    \n",
      "====================================================================================================\n",
      "Total params: 9,426,923\n",
      "Trainable params: 9,400,043\n",
      "Non-trainable params: 26,880\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "label_vae = Model(inputs=[x_in, y_in], outputs=[_x_output, _y_output])\n",
    "optimizer = Adam(lr=learning_rate, decay=decay)\n",
    "label_vae.compile(optimizer=optimizer, loss=[labeled_vae_loss, cls_loss])\n",
    "label_vae.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_1 (InputLayer)             (200, 28, 28, 1)      0                                            \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)                (200, 28, 28, 32)     320         input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "activation_5 (Activation)        (200, 28, 28, 32)     0           conv2d_1[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)                (200, 26, 26, 32)     9248        activation_5[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_1 (Conv2D)              (200, 28, 28, 64)     640         input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "activation_6 (Activation)        (200, 26, 26, 32)     0           conv2d_2[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_1 (BatchNormalization)    (200, 28, 28, 64)     256         enc_conv_1[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)   (200, 13, 13, 32)     0           activation_6[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "activation_1 (Activation)        (200, 28, 28, 64)     0           enc_bn_1[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)              (200, 13, 13, 32)     0           max_pooling2d_1[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_2 (Conv2D)              (200, 28, 28, 64)     36928       activation_1[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)                (200, 13, 13, 64)     18496       dropout_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_2 (BatchNormalization)    (200, 28, 28, 64)     256         enc_conv_2[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "activation_7 (Activation)        (200, 13, 13, 64)     0           conv2d_3[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "activation_2 (Activation)        (200, 28, 28, 64)     0           enc_bn_2[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)                (200, 11, 11, 64)     36928       activation_7[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_3 (Conv2D)              (200, 14, 14, 64)     36928       activation_2[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "activation_8 (Activation)        (200, 11, 11, 64)     0           conv2d_4[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_3 (BatchNormalization)    (200, 14, 14, 64)     256         enc_conv_3[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2D)   (200, 5, 5, 64)       0           activation_8[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "activation_3 (Activation)        (200, 14, 14, 64)     0           enc_bn_3[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "dropout_3 (Dropout)              (200, 5, 5, 64)       0           max_pooling2d_2[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "flatten_1 (Flatten)              (200, 12544)          0           activation_3[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "flatten_2 (Flatten)              (200, 1600)           0           dropout_3[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_4 (Dense)               (200, 512)            6423040     flatten_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_3 (Dense)                  (200, 512)            819712      flatten_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_4 (BatchNormalization)    (200, 512)            2048        enc_conv_4[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "activation_9 (Activation)        (200, 512)            0           dense_3[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "activation_4 (Activation)        (200, 512)            0           enc_bn_4[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "dropout_4 (Dropout)              (200, 512)            0           activation_9[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)              (200, 512)            0           activation_4[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dense_4 (Dense)                  (200, 10)             5130        dropout_4[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_1 (Dense)                  (200, 128)            65664       dropout_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_2 (Dense)                  (200, 128)            65664       dropout_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "activation_10 (Activation)       (200, 10)             0           dense_4[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "lambda_1 (Lambda)                (200, 128)            0           dense_1[0][0]                    \n",
      "                                                                   dense_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "concatenate_2 (Concatenate)      (200, 138)            0           activation_10[0][0]              \n",
      "                                                                   lambda_1[0][0]                   \n",
      "____________________________________________________________________________________________________\n",
      "enc_conv_10 (Dense)              (200, 12544)          1743616     concatenate_2[0][0]              \n",
      "____________________________________________________________________________________________________\n",
      "enc_bn_10 (BatchNormalization)   (200, 12544)          50176       enc_conv_10[1][0]                \n",
      "____________________________________________________________________________________________________\n",
      "activation_11 (Activation)       (200, 12544)          0           enc_bn_10[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dropout_5 (Dropout)              (200, 12544)          0           activation_11[1][0]              \n",
      "____________________________________________________________________________________________________\n",
      "reshape_1 (Reshape)              (200, 14, 14, 64)     0           dropout_5[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dec_trans_conv_11 (Conv2DTranspo (200, 14, 14, 64)     36928       reshape_1[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dec_bn_11 (BatchNormalization)   (200, 14, 14, 64)     256         dec_trans_conv_11[1][0]          \n",
      "____________________________________________________________________________________________________\n",
      "activation_12 (Activation)       (200, 14, 14, 64)     0           dec_bn_11[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dec_trans_conv_12 (Conv2DTranspo (200, 14, 14, 64)     36928       activation_12[1][0]              \n",
      "____________________________________________________________________________________________________\n",
      "dec_bn_12 (BatchNormalization)   (200, 14, 14, 64)     256         dec_trans_conv_12[1][0]          \n",
      "____________________________________________________________________________________________________\n",
      "activation_13 (Activation)       (200, 14, 14, 64)     0           dec_bn_12[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dec_trans_conv_13 (Conv2DTranspo (200, 28, 28, 64)     36928       activation_13[1][0]              \n",
      "____________________________________________________________________________________________________\n",
      "dec_bn_13 (BatchNormalization)   (200, 28, 28, 64)     256         dec_trans_conv_13[1][0]          \n",
      "____________________________________________________________________________________________________\n",
      "activation_14 (Activation)       (200, 28, 28, 64)     0           dec_bn_13[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "x_decoded (Conv2DTranspose)      (200, 28, 28, 1)      65          activation_14[1][0]              \n",
      "====================================================================================================\n",
      "Total params: 9,426,923\n",
      "Trainable params: 9,400,043\n",
      "Non-trainable params: 26,880\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "unlabeled_vae = Model(inputs=x_in, outputs=u_x_output)\n",
    "optimizer = Adam(lr=learning_rate, decay=decay)\n",
    "unlabeled_vae.compile(optimizer=optimizer, loss=unlabeled_vae_loss)\n",
    "unlabeled_vae.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from tqdm import tnrange, tqdm_notebook\n",
    "def fit_model(X_unlabeled, X_labeled, y_labeled, epochs):\n",
    "    assert len(X_unlabeled) % len(X_labeled) == 0, \\\n",
    "            (len(X_unlabeled), batch_size, len(X_labeled))\n",
    "    start = time.time()\n",
    "    history = []\n",
    "    \n",
    "    for epoch in tnrange(epochs):\n",
    "        unlabeled_index = np.arange(len(X_unlabeled))\n",
    "        np.random.shuffle(unlabeled_index)\n",
    "        \n",
    "        # Repeat the labeled data to match length of unlabeled data\n",
    "        labeled_index = []\n",
    "        for i in range(len(X_unlabeled) // len(X_labeled)):\n",
    "            l = np.arange(len(X_labeled))\n",
    "            np.random.shuffle(l)\n",
    "            labeled_index.append(l)\n",
    "        labeled_index = np.concatenate(labeled_index)\n",
    "        \n",
    "        batches = len(X_unlabeled) // batch_size\n",
    "        with tnrange(batches, leave=False) as pbar:\n",
    "            for i in range(batches):\n",
    "                # Labeled\n",
    "                index_range =  labeled_index[i * batch_size:(i+1) * batch_size]\n",
    "                loss = label_vae.train_on_batch([X_labeled[index_range], y_labeled[index_range]], \n",
    "                                                [X_labeled[index_range], y_labeled[index_range]])\n",
    "                \n",
    "                # Unlabeled\n",
    "                index_range =  unlabeled_index[i * batch_size:(i+1) * batch_size]\n",
    "                loss += [unlabeled_vae.train_on_batch(X_unlabeled[index_range],  X_unlabeled[index_range])]\n",
    "                pbar.desc = 'Loss: %s' % loss\n",
    "                history.append(loss)\n",
    "                \n",
    "                pbar.update(1)\n",
    "    \n",
    "   \n",
    "    done = time.time()\n",
    "    elapsed = done - start\n",
    "    print(\"Elapsed: \", elapsed)\n",
    "    \n",
    "    return history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting with sample_size: 1000\n",
      "\n",
      "Elapsed:  626.3456034660339\n",
      " * Accuracy: 97.5 %\n",
      " * Fit time elapsed: 626.9s\n"
     ]
    }
   ],
   "source": [
    "sample_size = 1000\n",
    "start = time.time()\n",
    "print('Fitting with sample_size: {}'.format(sample_size))\n",
    "\n",
    "if sample_size < len(X_train):\n",
    "    sss = StratifiedShuffleSplit(n_splits=2, test_size=sample_size / len(X_train), random_state=0)\n",
    "    _, index = sss.split(X_train, y_train)\n",
    "    X, y = X_train[index[1]], y_train[index[1]]\n",
    "    X_others, _ = X_train[index[0]], y_train[index[0]]\n",
    "else:\n",
    "    X, y = X_train, y_train\n",
    "\n",
    "y = np_utils.to_categorical(y)\n",
    "\n",
    "history = fit_model(X_others, X, y, epochs=epochs)\n",
    "\n",
    "classifier = Model(inputs=[x_in], outputs=[_y_output])\n",
    "\n",
    "y_pred = np.argmax(classifier.predict(X_test, batch_size=batch_size), axis=-1)\n",
    "score = accuracy_score(y_test, y_pred)\n",
    "\n",
    "end = time.time()\n",
    "elapsed = end - start\n",
    "print(' * Accuracy: %.1f %%' % (100. * score))\n",
    "print(' * Fit time elapsed: %.1fs' % elapsed)\n",
    "\n",
    "results = []\n",
    "results.append({'sample_size': sample_size, 'accuracy': score, 'time': elapsed})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>labeled_loss</th>\n",
       "      <th>labeled_loss2</th>\n",
       "      <th>classifier_loss</th>\n",
       "      <th>unlabeled_loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2950.000000</td>\n",
       "      <td>2950.000000</td>\n",
       "      <td>2950.000000</td>\n",
       "      <td>2950.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>109.359922</td>\n",
       "      <td>104.140978</td>\n",
       "      <td>5.218944</td>\n",
       "      <td>108.953968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>49.704825</td>\n",
       "      <td>27.811763</td>\n",
       "      <td>23.087066</td>\n",
       "      <td>25.612387</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>89.658028</td>\n",
       "      <td>89.390045</td>\n",
       "      <td>0.028345</td>\n",
       "      <td>94.809906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0%</th>\n",
       "      <td>89.658028</td>\n",
       "      <td>89.390045</td>\n",
       "      <td>0.028345</td>\n",
       "      <td>94.809906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>95.304316</td>\n",
       "      <td>94.685501</td>\n",
       "      <td>0.417195</td>\n",
       "      <td>100.923922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>98.431107</td>\n",
       "      <td>97.386917</td>\n",
       "      <td>0.985630</td>\n",
       "      <td>103.080780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>105.460464</td>\n",
       "      <td>102.913551</td>\n",
       "      <td>2.603982</td>\n",
       "      <td>107.032776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95%</th>\n",
       "      <td>152.373417</td>\n",
       "      <td>137.385592</td>\n",
       "      <td>16.067925</td>\n",
       "      <td>137.611858</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99%</th>\n",
       "      <td>300.924666</td>\n",
       "      <td>215.205708</td>\n",
       "      <td>85.311641</td>\n",
       "      <td>209.207086</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>861.343933</td>\n",
       "      <td>629.905701</td>\n",
       "      <td>310.686188</td>\n",
       "      <td>549.917358</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       labeled_loss  labeled_loss2  classifier_loss  unlabeled_loss\n",
       "count   2950.000000    2950.000000      2950.000000     2950.000000\n",
       "mean     109.359922     104.140978         5.218944      108.953968\n",
       "std       49.704825      27.811763        23.087066       25.612387\n",
       "min       89.658028      89.390045         0.028345       94.809906\n",
       "0%        89.658028      89.390045         0.028345       94.809906\n",
       "25%       95.304316      94.685501         0.417195      100.923922\n",
       "50%       98.431107      97.386917         0.985630      103.080780\n",
       "75%      105.460464     102.913551         2.603982      107.032776\n",
       "95%      152.373417     137.385592        16.067925      137.611858\n",
       "99%      300.924666     215.205708        85.311641      209.207086\n",
       "max      861.343933     629.905701       310.686188      549.917358"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f277d8127b8>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFpCAYAAABTSWtMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2804760358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.DataFrame(history, columns=['labeled_loss', 'labeled_loss2', 'classifier_loss', 'unlabeled_loss'])\n",
    "display(df.describe(percentiles=[0.25 * i for i in range(4)] + [0.95, 0.99]))\n",
    "df.plot(figsize=(8, 6))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generator Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "g_y = Input(shape=(num_classes,))\n",
    "g_z = Input(shape=(latent_dim,))\n",
    "g_merged = concatenate([g_y, g_z])\n",
    "g_output = inst_layers(decoder_layers, g_merged)\n",
    "generator = Model([g_y, g_z], g_output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAksAAAJCCAYAAADQsoPKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\na4kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f277d27ee80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "n = 10\n",
    "digit_size = 28\n",
    "figure = np.zeros((digit_size * n, digit_size * n))\n",
    "\n",
    "for j in range(n):\n",
    "    for i in range(n):\n",
    "        y_sample = np.zeros(num_classes).reshape(1, num_classes)\n",
    "        y_sample[0, i] = 1\n",
    "        z_sample = np.random.normal(size=latent_dim).reshape(1, latent_dim)\n",
    "        x_decoded = generator.predict([y_sample, z_sample], batch_size=1)\n",
    "        digit = x_decoded.reshape(digit_size, digit_size)\n",
    "        \n",
    "        d_x = i * digit_size\n",
    "        d_y = j * digit_size\n",
    "        figure[d_x:d_x + digit_size, d_y:d_y + digit_size] = digit\n",
    "\n",
    "plt.figure(figsize=(10, 10))\n",
    "plt.imshow(figure)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>sample_size</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.9752</td>\n",
       "      <td>1000</td>\n",
       "      <td>626.904411</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   accuracy  sample_size        time\n",
       "0    0.9752         1000  626.904411"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.DataFrame(results)\n",
    "display(df)\n",
    "df.to_csv('vae-m2-mnist_results.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}