mnist_vae.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# what is the variational?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import os\n",
    "import time\n",
    "import random\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = '0'\n",
    "\n",
    "import tensorflow as tf\n",
    "import tensorflow.keras.backend as K\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.layers import *\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.losses import mse\n",
    "from tensorflow.keras.datasets import mnist\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "x_train, x_test = [x/255.0 for x in [x_train, x_test]]\n",
    "y_train, y_test = [tf.keras.utils.to_categorical(x) for x in [y_train, y_test]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "ZDIM = 4\n",
    "\n",
    "def get_encoder():\n",
    "  x = in1 = Input((28,28))\n",
    "  x = Reshape((28,28,1))(x)\n",
    "\n",
    "  x = Conv2D(64, (5,5), padding='same', strides=(2,2))(x)\n",
    "  x = BatchNormalization()(x)\n",
    "  x = ELU()(x)\n",
    "\n",
    "  x = Conv2D(128, (5,5), padding='same', strides=(2,2))(x)\n",
    "  x = BatchNormalization()(x)\n",
    "  x = ELU()(x)\n",
    "\n",
    "  x = Flatten()(x)\n",
    "  x = Dense(ZDIM)(x)\n",
    "  return Model(in1, x)\n",
    "\n",
    "def get_decoder():\n",
    "  x = in1 = Input((ZDIM,))\n",
    "\n",
    "  x = Dense(7*7*64)(x)\n",
    "  x = BatchNormalization()(x)\n",
    "  x = ELU()(x)\n",
    "  x = Reshape((7,7,64))(x)\n",
    "\n",
    "  x = Conv2DTranspose(128, (5,5), strides=(2,2), padding='same')(x)\n",
    "  x = BatchNormalization()(x)\n",
    "  x = ELU()(x)\n",
    "\n",
    "  x = Conv2DTranspose(1, (5,5), strides=(2,2), padding='same')(x)\n",
    "  x = Activation('sigmoid')(x)\n",
    "  x = Reshape((28,28))(x)\n",
    "  return Model(in1, x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 28, 28)]          0         \n",
      "_________________________________________________________________\n",
      "reshape (Reshape)            (None, 28, 28, 1)         0         \n",
      "_________________________________________________________________\n",
      "conv2d (Conv2D)              (None, 14, 14, 64)        1664      \n",
      "_________________________________________________________________\n",
      "batch_normalization (BatchNo (None, 14, 14, 64)        256       \n",
      "_________________________________________________________________\n",
      "elu (ELU)                    (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 7, 7, 128)         204928    \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 7, 7, 128)         512       \n",
      "_________________________________________________________________\n",
      "elu_1 (ELU)                  (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 6272)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 4)                 25092     \n",
      "_________________________________________________________________\n",
      "model_1 (Model)              (None, 28, 28)            236865    \n",
      "=================================================================\n",
      "Total params: 469,317\n",
      "Trainable params: 462,405\n",
      "Non-trainable params: 6,912\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/10\n",
      "60000/60000 [==============================] - 12s 201us/sample - loss: 0.0419 - val_loss: 0.0374\n",
      "Epoch 2/10\n",
      "60000/60000 [==============================] - 6s 104us/sample - loss: 0.0349 - val_loss: 0.0338\n",
      "Epoch 3/10\n",
      "60000/60000 [==============================] - 6s 107us/sample - loss: 0.0338 - val_loss: 0.0331\n",
      "Epoch 4/10\n",
      "60000/60000 [==============================] - 6s 107us/sample - loss: 0.0330 - val_loss: 0.0329\n",
      "Epoch 5/10\n",
      "60000/60000 [==============================] - 6s 107us/sample - loss: 0.0325 - val_loss: 0.0312\n",
      "Epoch 6/10\n",
      "60000/60000 [==============================] - 6s 104us/sample - loss: 0.0321 - val_loss: 0.0310\n",
      "Epoch 7/10\n",
      "60000/60000 [==============================] - 6s 107us/sample - loss: 0.0317 - val_loss: 0.0307\n",
      "Epoch 8/10\n",
      "60000/60000 [==============================] - 6s 105us/sample - loss: 0.0315 - val_loss: 0.0306\n",
      "Epoch 9/10\n",
      "60000/60000 [==============================] - 6s 108us/sample - loss: 0.0312 - val_loss: 0.0301\n",
      "Epoch 10/10\n",
      "60000/60000 [==============================] - 6s 108us/sample - loss: 0.0309 - val_loss: 0.0300\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7fc27db12dd8>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# normal autoencoder without the variational\n",
    "enc = get_encoder()\n",
    "ae = Model(enc.input, get_decoder()(enc.output))\n",
    "ae.compile('adam', 'mse')\n",
    "ae.summary()\n",
    "ae.fit(x_train, x_train, batch_size=128, epochs=10, shuffle=1, validation_data=(x_test, x_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fc36088b550>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# while the autoencoder might work without the variational, the sampling doesn't\n",
    "figsize(16,16)\n",
    "imshow(np.concatenate(ae.predict(np.array([random.choice(x_test) for i in range(10)])), axis=1))\n",
    "figure()\n",
    "imshow(np.concatenate(ae.layers[-1].predict(np.random.normal(size=(10, ZDIM))), axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Output model_4 missing from loss dictionary. We assume this was done on purpose. The fit and evaluate APIs will not be expecting any data to be passed to model_4.\n"
     ]
    }
   ],
   "source": [
    "in1 = Input((28,28))\n",
    "x = get_encoder()(in1)\n",
    "\n",
    "# add the variational\n",
    "z_mu = Dense(ZDIM)(x)\n",
    "z_log_var = Dense(ZDIM)(x)\n",
    "z = Lambda(lambda x: x[0] + K.exp(0.5 * x[1]) * K.random_normal(shape=K.shape(x[0])))([z_mu, z_log_var])\n",
    "out = get_decoder()(z)\n",
    "\n",
    "mse_loss = mse(Reshape((28*28,))(in1), Reshape((28*28,))(out)) * 28 * 28\n",
    "kl_loss = 1 + z_log_var - K.square(z_mu) - K.exp(z_log_var)\n",
    "kl_loss = -0.5 * K.mean(kl_loss, axis=-1)\n",
    "\n",
    "vae = Model(in1, out)\n",
    "vae.add_loss(K.mean(mse_loss + kl_loss))\n",
    "vae.compile('adam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([[-0.36135393,  1.8782282 , -2.056991  , -0.60052466]],\n",
       "       dtype=float32),\n",
       " array([[ 0.01278172,  0.00325583, -0.00295223,  0.02152798]],\n",
       "       dtype=float32),\n",
       " array([[ 0.01860134, -0.03426083, -0.00210588, -0.04922588]],\n",
       "       dtype=float32)]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# z is sampled from z_mu and z_log_var with gaussian noise\n",
    "test = Model(in1, [z, z_mu, z_log_var])\n",
    "test.predict(x_train[0:1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/10\n",
      "60000/60000 [==============================] - 10s 172us/sample - loss: 106.9424 - val_loss: 49.5300\n",
      "Epoch 2/10\n",
      "60000/60000 [==============================] - 7s 123us/sample - loss: 40.6906 - val_loss: 38.4624\n",
      "Epoch 3/10\n",
      "60000/60000 [==============================] - 8s 126us/sample - loss: 37.5029 - val_loss: 36.0076\n",
      "Epoch 4/10\n",
      "60000/60000 [==============================] - 7s 123us/sample - loss: 35.5239 - val_loss: 34.0764\n",
      "Epoch 5/10\n",
      "60000/60000 [==============================] - 7s 124us/sample - loss: 34.0624 - val_loss: 32.7335\n",
      "Epoch 6/10\n",
      "60000/60000 [==============================] - 8s 125us/sample - loss: 33.0450 - val_loss: 32.2324\n",
      "Epoch 7/10\n",
      "60000/60000 [==============================] - 7s 123us/sample - loss: 32.2560 - val_loss: 31.2762\n",
      "Epoch 8/10\n",
      "60000/60000 [==============================] - 7s 123us/sample - loss: 31.7095 - val_loss: 30.9809\n",
      "Epoch 9/10\n",
      "60000/60000 [==============================] - 7s 125us/sample - loss: 31.1592 - val_loss: 30.4665\n",
      "Epoch 10/10\n",
      "60000/60000 [==============================] - 7s 123us/sample - loss: 30.7849 - val_loss: 29.9908\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7fc3544e6a90>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vae.fit(x_train, batch_size=128, epochs=10, shuffle=1, validation_data=(x_test, None))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fc354523cf8>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(16,16)\n",
    "imshow(np.concatenate(vae.predict(np.array([random.choice(x_test) for i in range(10)])), axis=1))\n",
    "figure()\n",
    "imshow(np.concatenate(vae.layers[5].predict(np.random.normal(size=(10, ZDIM))), axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}