CBNGAN_mnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import division\n",
    "import os\n",
    "import time\n",
    "import math\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from IPython.display import clear_output\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Layer, InputSpec, Conv2D, Conv2DTranspose, Activation, Reshape, BatchNormalization, AveragePooling2D\n",
    "from tensorflow.keras.layers import Input, UpSampling2D, Dropout, Concatenate, add, Dense, Multiply, LeakyReLU, ReLU, Flatten, Dot, Embedding\n",
    "from tensorflow.keras import initializers, regularizers, constraints, Model, Sequential\n",
    "from tensorflow.keras import backend as K\n",
    "from tensorflow.keras.utils import Sequence\n",
    "from tensorflow.python.keras.utils import conv_utils\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, LearningRateScheduler, TensorBoard, EarlyStopping\n",
    "from tensorflow.keras.optimizers import Adam, SGD\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.initializers import RandomNormal\n",
    "from tensorflow.keras.losses import BinaryCrossentropy\n",
    "from tensorflow.keras.utils import to_categorical, plot_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from csv file\n",
    "train = pd.read_csv(\"../../mnist/data/train.csv\")\n",
    "test = pd.read_csv(\"../../mnist/data/test.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   label  pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  \\\n",
       "0      1       0       0       0       0       0       0       0       0   \n",
       "1      0       0       0       0       0       0       0       0       0   \n",
       "2      1       0       0       0       0       0       0       0       0   \n",
       "3      4       0       0       0       0       0       0       0       0   \n",
       "4      0       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel8  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0         0         0         0   \n",
       "3       0  ...         0         0         0         0         0         0   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel780  pixel781  pixel782  pixel783  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 785 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = train['label']\n",
    "train_data = train.drop(\"label\", axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(42000, 784)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = (train_data.values.reshape(-1, 28, 28, 1) - 127.5) / 127.5\n",
    "#train_data = train_data.values / 255."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]]],\n",
       "\n",
       "\n",
       "       [[[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]]],\n",
       "\n",
       "\n",
       "       [[[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]]],\n",
       "\n",
       "\n",
       "       ...,\n",
       "\n",
       "\n",
       "       [[[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]]],\n",
       "\n",
       "\n",
       "       [[[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]]],\n",
       "\n",
       "\n",
       "       [[[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]],\n",
       "\n",
       "        [[-1.],\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         ...,\n",
       "         [-1.],\n",
       "         [-1.],\n",
       "         [-1.]]]])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(42000, 28, 28, 1)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = np.sort(labels.unique())\n",
    "oh_labels = to_categorical(labels, num_classes=len(classes))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = labels.values.reshape((-1, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(42000, 10)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oh_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(42000, 1)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "OUTPUT_PATH = 'outputs'\n",
    "TRAIN_LOGDIR = os.path.join(\"logs\", \"tensorflow\", 'train_data') # Sets up a log directory.\n",
    "#TEST_LOGDIR = os.path.join(\"logs\", \"tensorflow\", 'test_data') # Sets up a log directory.\n",
    "if not os.path.exists(OUTPUT_PATH):\n",
    "    os.makedirs(OUTPUT_PATH)\n",
    "\n",
    "BATCH_SIZE = 64\n",
    "NOISE_DIM = 100\n",
    "\n",
    "ITERATIONS = 100000\n",
    "DECAY_ITERATIONS = 80000\n",
    "CURRENT_ITERATION = 1 # Epoch start from\n",
    "SAVE_EVERY_N_ITER = 5000 # Save checkpoint at every n epoch\n",
    "\n",
    "D_G_TRAINING_RATIO = 5\n",
    "LR = 0.0002\n",
    "BETA_1 = 0.0\n",
    "BETA_2 = 0.9\n",
    "\n",
    "MIN_LR = 0.000001\n",
    "DECAY_FACTOR=1.00004\n",
    "\n",
    "# Creates a file writer for the log directory.\n",
    "file_writer = tf.summary.create_file_writer(TRAIN_LOGDIR)\n",
    "#test_file_writer = tf.summary.create_file_writer(TEST_LOGDIR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ConditionalBatchNormalization(Model):\n",
    "    \"\"\"\n",
    "    https://github.com/crcrpar/pytorch.sngan_projection/blob/master/links/conditional_batchnorm.py\n",
    "    Conditional Batch Normalization Base Class\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, \n",
    "               num_classes,\n",
    "               units,\n",
    "               axis=-1,\n",
    "               momentum=0.99,\n",
    "               epsilon=1e-3,\n",
    "               center=False,\n",
    "               scale=False,\n",
    "               beta_initializer='zeros',\n",
    "               gamma_initializer='ones',\n",
    "               moving_mean_initializer='zeros',\n",
    "               moving_variance_initializer='ones',\n",
    "               beta_regularizer=None,\n",
    "               gamma_regularizer=None,\n",
    "               beta_constraint=None,\n",
    "               gamma_constraint=None,\n",
    "               renorm=False,\n",
    "               renorm_clipping=None,\n",
    "               renorm_momentum=0.99,\n",
    "               fused=None,\n",
    "               trainable=True,\n",
    "               virtual_batch_size=None,\n",
    "               adjustment=None,\n",
    "               name=None,\n",
    "               **kwargs):\n",
    "        super(ConditionalBatchNormalization, self).__init__(name, **kwargs)\n",
    "        self.weight_embedding = Embedding(input_dim=num_classes, output_dim=units, input_length=1, embeddings_initializer='ones')\n",
    "        self.bias_embedding = Embedding(input_dim=num_classes, output_dim=units, input_length=1, embeddings_initializer='zeros')\n",
    "        self.bn = BatchNormalization(axis=-1,\n",
    "               momentum=0.99,\n",
    "               epsilon=1e-3,\n",
    "               center=False,\n",
    "               scale=False,\n",
    "               beta_initializer='zeros',\n",
    "               gamma_initializer='ones',\n",
    "               moving_mean_initializer='zeros',\n",
    "               moving_variance_initializer='ones',\n",
    "               beta_regularizer=None,\n",
    "               gamma_regularizer=None,\n",
    "               beta_constraint=None,\n",
    "               gamma_constraint=None,\n",
    "               renorm=False,\n",
    "               renorm_clipping=None,\n",
    "               renorm_momentum=0.99,\n",
    "               fused=None,\n",
    "               trainable=True,\n",
    "               virtual_batch_size=None,\n",
    "               adjustment=None,)\n",
    "\n",
    "    def call(self, inputs, training=None):\n",
    "        \"\"\"Call `Layer`\"\"\"\n",
    "        \n",
    "        class_weights = self.weight_embedding(inputs[1])\n",
    "        class_bias = self.bias_embedding(inputs[1])\n",
    "\n",
    "        if len(class_weights.shape) == 3:\n",
    "            class_weights = tf.squeeze(class_weights, [1])\n",
    "        if len(class_bias.shape) == 3:\n",
    "            class_bias =tf.squeeze(class_bias, [1])\n",
    "        \n",
    "        output = self.bn(inputs[0], training=training)\n",
    "        \n",
    "        if len(class_weights.shape) == 1:\n",
    "            class_weights = tf.expand_dims(class_weights, 0)\n",
    "        if len(class_bias.shape) == 1:\n",
    "            class_bias = tf.expand_dims(class_bias, 0)\n",
    "            \n",
    "        \n",
    "        class_weights = class_weights[:, tf.newaxis, tf.newaxis, :]\n",
    "        class_bias = class_bias[:, tf.newaxis, tf.newaxis, :]\n",
    "        output = class_weights * output + class_bias\n",
    "        return output\n",
    "    \n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return tensor_shape.TensorShape(input_shape[0])\n",
    "\n",
    "class SpectralNormalization(tf.keras.layers.Wrapper):\n",
    "    \"\"\"\n",
    "    Reference from WeightNormalization implementation of TF Addons\n",
    "    SpectralNormalization wrapper works for keras CNN and Dense (RNN not tested).\n",
    "    ```python\n",
    "      net = SpectralNormalization(\n",
    "          tf.keras.layers.Conv2D(2, 2, activation='relu'),\n",
    "          input_shape=(32, 32, 3))(x)\n",
    "    ```\n",
    "    Arguments:\n",
    "      layer: a layer instance.\n",
    "    Raises:\n",
    "      ValueError: If `Layer` does not contain a `kernel` of weights\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, layer, power_iter=1, **kwargs):\n",
    "        super(SpectralNormalization, self).__init__(layer, **kwargs)\n",
    "        self._track_trackable(layer, name='layer')\n",
    "        self.power_iter = 1\n",
    "        self.is_rnn = isinstance(self.layer, tf.keras.layers.RNN)\n",
    "        self.is_embedding = isinstance(self.layer, tf.keras.layers.Embedding)\n",
    "\n",
    "    def build(self, input_shape):\n",
    "        \"\"\"Build `Layer`\"\"\"\n",
    "        input_shape = tf.TensorShape(input_shape)\n",
    "        self.input_spec = tf.keras.layers.InputSpec(\n",
    "            shape=[None] + input_shape[1:])\n",
    "\n",
    "        if not self.layer.built:\n",
    "            self.layer.build(input_shape)\n",
    "\n",
    "        if self.is_rnn:\n",
    "            kernel_layer = self.layer.cell\n",
    "        else:\n",
    "            kernel_layer = self.layer\n",
    "\n",
    "        if not hasattr(kernel_layer, 'kernel') and not hasattr(kernel_layer, 'embeddings'):\n",
    "            raise ValueError('`SpectralNormalization` must wrap a layer that'\n",
    "                             ' contains a `kernel or embedding` for weights')\n",
    "\n",
    "        if self.is_rnn:\n",
    "            kernel = kernel_layer.recurrent_kernel\n",
    "        elif self.is_embedding:\n",
    "            kernel = kernel_layer.embeddings\n",
    "        else:\n",
    "            kernel = kernel_layer.kernel\n",
    "            \n",
    "        self.kernel_shape = kernel.shape\n",
    "        \n",
    "        self.u = self.add_weight(name='u',\n",
    "                             shape=(self.kernel_shape[-1], 1),\n",
    "                             initializer='random_normal',\n",
    "                             trainable=False)\n",
    "        self.w = kernel\n",
    "        self.built = True\n",
    "    \n",
    "    def call(self, inputs, training=True):\n",
    "        \"\"\"Call `Layer`\"\"\"\n",
    "\n",
    "        with tf.name_scope('compute_weights'):\n",
    "            \n",
    "            weight = tf.reshape(self.w, shape=[self.kernel_shape[-1], -1])\n",
    "            # power iteration\n",
    "            for i in range(self.power_iter):\n",
    "                v = K.l2_normalize(tf.matmul(tf.transpose(weight), self.u))\n",
    "                u = K.l2_normalize(tf.matmul(weight, v))\n",
    "            v = tf.stop_gradient(v)\n",
    "            u = tf.stop_gradient(u)\n",
    "            \n",
    "            sigma = tf.matmul(tf.matmul(tf.transpose(u), weight), v)\n",
    "            kernel = self.w / sigma\n",
    "\n",
    "            if self.is_rnn:\n",
    "                print(self.is_rnn)\n",
    "                self.layer.cell.recurrent_kernel = kernel\n",
    "                update_kernel = tf.identity(self.layer.cell.recurrent_kernel)\n",
    "            else:\n",
    "                self.layer.kernel = kernel\n",
    "                update_kernel = tf.identity(self.layer.kernel)\n",
    "\n",
    "            # Ensure we calculate result after updating kernel.\n",
    "            with tf.control_dependencies([update_kernel]):\n",
    "                outputs = self.layer(inputs)\n",
    "                return outputs\n",
    "\n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return tf.TensorShape(\n",
    "            self.layer.compute_output_shape(input_shape).as_list())\n",
    "    \n",
    "class GlobalSumPooling2D(Layer):\n",
    "    def __init__(self, data_format='channels_last', **kwargs):\n",
    "        super(GlobalSumPooling2D, self).__init__(**kwargs)\n",
    "        self.input_spec = InputSpec(ndim=4)\n",
    "        self.data_format = conv_utils.normalize_data_format(data_format)\n",
    "    \n",
    "    def call(self, inputs):\n",
    "        if self.data_format == 'channels_last':\n",
    "            return K.sum(inputs, axis=[1, 2])\n",
    "        else:\n",
    "            return K.sum(inputs, axis=[2, 3])\n",
    "    \n",
    "    def get_config(self):\n",
    "        config = {'data_format': self.data_format}\n",
    "        base_config = super(GlobalSumPooling2D, self).get_config()\n",
    "        return dict(list(base_config.items()) + list(config.items()))\n",
    "    \n",
    "    def compute_output_shape(self, input_shape):\n",
    "        input_shape = tensor_shape.TensorShape(input_shape).as_list()\n",
    "        if self.data_format == 'channels_last':\n",
    "            return tensor_shape.TensorShape([input_shape[0], input_shape[3]])\n",
    "        else:\n",
    "            return tensor_shape.TensorShape([input_shape[0], input_shape[1]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def upsample_residual_block(inputs, in_filters, filters, kernel_size=3, strides=1, padding='same',\n",
    "                            activation=tf.nn.leaky_relu, upsample=True, name=''):\n",
    "    '''\n",
    "        Upsampling Residual Block\n",
    "    '''\n",
    "    kernel_initializer = 'glorot_uniform'\n",
    "    x_input, class_input = inputs\n",
    "    \n",
    "    x = ConditionalBatchNormalization(num_classes=len(classes), units=in_filters)([x_input, class_input])\n",
    "    x = Activation(activation)(x)\n",
    "    if upsample:\n",
    "        x = UpSampling2D(size=2, interpolation='nearest')(x)\n",
    "    x = Conv2D(filters, kernel_size, strides, padding=padding,\n",
    "                   kernel_initializer=kernel_initializer)(x)\n",
    "    \n",
    "    x = ConditionalBatchNormalization(num_classes=len(classes), units=filters)([x, class_input])\n",
    "    x = Activation(activation)(x)\n",
    "    x = Conv2D(filters, kernel_size, strides, padding=padding, kernel_initializer=kernel_initializer)(x)\n",
    "    \n",
    "    if in_filters != filters or upsample:\n",
    "        if upsample:\n",
    "            x_input = UpSampling2D(size=2, interpolation='nearest')(x_input)\n",
    "        x_input = Conv2D(filters, 1, 1, padding='valid',\n",
    "                   kernel_initializer=kernel_initializer)(x_input)\n",
    "    \n",
    "    x += x_input\n",
    "    return x\n",
    "\n",
    "def initial_downsample_block(inputs, in_filters, filters, kernel_size=3, strides=1, padding='same',\n",
    "                              activation=tf.nn.leaky_relu, name=''):\n",
    "    '''\n",
    "        Downsampling Residual Block\n",
    "    '''\n",
    "    kernel_initializer = 'glorot_uniform'\n",
    "    x_input = inputs\n",
    "\n",
    "    x = SpectralNormalization(Conv2D(filters, kernel_size, strides, padding=padding,\n",
    "                   kernel_initializer=kernel_initializer))(x_input)\n",
    "    x = Activation(activation)(x)\n",
    "    x = SpectralNormalization(Conv2D(filters, kernel_size, strides, padding=padding, kernel_initializer=kernel_initializer))(x)\n",
    "    \n",
    "    x = AveragePooling2D(pool_size=2)(x)\n",
    "    \n",
    "    x_input = AveragePooling2D(pool_size=2)(x_input)\n",
    "    x_input = SpectralNormalization(Conv2D(filters, 1, 1, padding='valid',\n",
    "                   kernel_initializer=kernel_initializer))(x_input)\n",
    "    \n",
    "    x += x_input\n",
    "    return x\n",
    "\n",
    "def downsample_residual_block(inputs, in_filters, filters, kernel_size=3, strides=1, padding='same',\n",
    "                              activation=tf.nn.leaky_relu, downsample=True, name=''):\n",
    "    '''\n",
    "        Downsampling Residual Block\n",
    "    '''\n",
    "    kernel_initializer = 'glorot_uniform'\n",
    "    x_input = inputs\n",
    "\n",
    "    x = Activation(activation)(x_input)\n",
    "    x = SpectralNormalization(Conv2D(filters, kernel_size, strides, padding=padding,\n",
    "                   kernel_initializer=kernel_initializer))(x)\n",
    "    \n",
    "    x = Activation(activation)(x)\n",
    "    x = SpectralNormalization(Conv2D(filters, kernel_size, strides, padding=padding, kernel_initializer=kernel_initializer))(x)\n",
    "    \n",
    "    if downsample:\n",
    "        x = AveragePooling2D(pool_size=2)(x)\n",
    "    \n",
    "    if in_filters != filters or downsample:\n",
    "        x_input = SpectralNormalization(Conv2D(filters, 1, 1, padding='valid',\n",
    "                   kernel_initializer=kernel_initializer))(x_input)\n",
    "        if downsample:\n",
    "            x_input = AveragePooling2D(pool_size=2)(x_input)\n",
    "\n",
    "    x += x_input\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# activation = tf.nn.leaky_relu\n",
    "activation = tf.keras.activations.relu\n",
    "\n",
    "def CGAN_generator(input_z_shape=NOISE_DIM, input_y_shape=(1)):\n",
    "    input_z_layer = Input(input_z_shape)\n",
    "    input_y_layer = Input(input_y_shape)\n",
    "    \n",
    "    z = Dense(7*7*256, use_bias=False)(input_z_layer)\n",
    "    z =Reshape((7, 7, 256))(z)\n",
    "    x = z\n",
    "    \n",
    "    x = upsample_residual_block([x, input_y_layer], in_filters=256, filters=128, kernel_size=3, strides=1, activation=activation, upsample=True)\n",
    "    x = upsample_residual_block([x, input_y_layer], in_filters=128, filters=64, kernel_size=3, strides=1, activation=activation, upsample=True)\n",
    "    \n",
    "    x = BatchNormalization()(x)\n",
    "    x = Activation(activation)(x)\n",
    "\n",
    "    output = Conv2D(1, (3, 3), strides=(1, 1), padding='same', use_bias=False, activation=\"tanh\")(x)\n",
    "    \n",
    "    model = Model(inputs=[input_z_layer, input_y_layer], outputs=output)\n",
    "    return model\n",
    "\n",
    "\n",
    "def CGAN_discriminator(input_x_shape=train_data.shape[1:], input_y_shape=(1)):\n",
    "    input_x_layer = Input(input_x_shape)\n",
    "    input_y_layer = Input(input_y_shape)\n",
    "    \n",
    "    x = initial_downsample_block(input_x_layer, in_filters=3, filters=64, kernel_size=3, strides=1, activation=activation)\n",
    "    x = downsample_residual_block(x, in_filters=64, filters=128, kernel_size=3, strides=1, activation=activation, downsample=True)\n",
    "    \n",
    "    pooling_out = GlobalSumPooling2D()(x)\n",
    "    \n",
    "    embed = SpectralNormalization(Embedding(len(classes), 128, embeddings_initializer='glorot_uniform', input_length=1))(input_y_layer)\n",
    "    \n",
    "    embed = tf.reshape(embed, [-1, 128])\n",
    "    projection = tf.reduce_sum(pooling_out * embed, axis=1, keepdims=True)\n",
    "    \n",
    "    linear_output = SpectralNormalization(Dense(1, kernel_initializer='glorot_uniform'))(pooling_out)\n",
    "    output = linear_output + projection\n",
    "    \n",
    "    model = Model(inputs=[input_x_layer, input_y_layer], outputs=output)\n",
    "    return model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "generator = CGAN_generator()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            [(None, 100)]        0                                            \n",
      "__________________________________________________________________________________________________\n",
      "dense (Dense)                   (None, 12544)        1254400     input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "reshape (Reshape)               (None, 7, 7, 256)    0           dense[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "input_2 (InputLayer)            [(None, 1)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conditional_batch_normalization (None, 7, 7, 256)    5632        reshape[0][0]                    \n",
      "                                                                 input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "activation (Activation)         (None, 7, 7, 256)    0           conditional_batch_normalization[0\n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d (UpSampling2D)    (None, 14, 14, 256)  0           activation[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv2d (Conv2D)                 (None, 14, 14, 128)  295040      up_sampling2d[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conditional_batch_normalization (None, 14, 14, 128)  2816        conv2d[0][0]                     \n",
      "                                                                 input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "activation_1 (Activation)       (None, 14, 14, 128)  0           conditional_batch_normalization_1\n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d_1 (UpSampling2D)  (None, 14, 14, 256)  0           reshape[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)               (None, 14, 14, 128)  147584      activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)               (None, 14, 14, 128)  32896       up_sampling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_add (TensorFlowOpLa [(None, 14, 14, 128) 0           conv2d_1[0][0]                   \n",
      "                                                                 conv2d_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conditional_batch_normalization (None, 14, 14, 128)  2816        tf_op_layer_add[0][0]            \n",
      "                                                                 input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "activation_2 (Activation)       (None, 14, 14, 128)  0           conditional_batch_normalization_2\n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d_2 (UpSampling2D)  (None, 28, 28, 128)  0           activation_2[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)               (None, 28, 28, 64)   73792       up_sampling2d_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conditional_batch_normalization (None, 28, 28, 64)   1408        conv2d_3[0][0]                   \n",
      "                                                                 input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "activation_3 (Activation)       (None, 28, 28, 64)   0           conditional_batch_normalization_3\n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d_3 (UpSampling2D)  (None, 28, 28, 128)  0           tf_op_layer_add[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)               (None, 28, 28, 64)   36928       activation_3[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_5 (Conv2D)               (None, 28, 28, 64)   8256        up_sampling2d_3[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_add_1 (TensorFlowOp [(None, 28, 28, 64)] 0           conv2d_4[0][0]                   \n",
      "                                                                 conv2d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_4 (BatchNor (None, 28, 28, 64)   256         tf_op_layer_add_1[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "activation_4 (Activation)       (None, 28, 28, 64)   0           batch_normalization_4[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_6 (Conv2D)               (None, 28, 28, 1)    576         activation_4[0][0]               \n",
      "==================================================================================================\n",
      "Total params: 1,862,400\n",
      "Trainable params: 1,861,120\n",
      "Non-trainable params: 1,280\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "generator.summary()\n",
    "plot_model(generator, show_shapes=True, dpi=64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "discriminator = CGAN_discriminator()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_3 (InputLayer)            [(None, 28, 28, 1)]  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization (Spectra (None, 28, 28, 64)   704         input_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "activation_5 (Activation)       (None, 28, 28, 64)   0           spectral_normalization[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_1 (Spect (None, 28, 28, 64)   36992       activation_5[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_1 (AveragePoo (None, 14, 14, 1)    0           input_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d (AveragePooli (None, 14, 14, 64)   0           spectral_normalization_1[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_2 (Spect (None, 14, 14, 64)   192         average_pooling2d_1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_add_2 (TensorFlowOp [(None, 14, 14, 64)] 0           average_pooling2d[0][0]          \n",
      "                                                                 spectral_normalization_2[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "activation_6 (Activation)       (None, 14, 14, 64)   0           tf_op_layer_add_2[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_3 (Spect (None, 14, 14, 128)  73984       activation_6[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_7 (Activation)       (None, 14, 14, 128)  0           spectral_normalization_3[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_4 (Spect (None, 14, 14, 128)  147712      activation_7[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_5 (Spect (None, 14, 14, 128)  8448        tf_op_layer_add_2[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_2 (AveragePoo (None, 7, 7, 128)    0           spectral_normalization_4[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_3 (AveragePoo (None, 7, 7, 128)    0           spectral_normalization_5[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "input_4 (InputLayer)            [(None, 1)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_add_3 (TensorFlowOp [(None, 7, 7, 128)]  0           average_pooling2d_2[0][0]        \n",
      "                                                                 average_pooling2d_3[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_6 (Spect (None, 1, 128)       1408        input_4[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "global_sum_pooling2d (GlobalSum (None, 128)          0           tf_op_layer_add_3[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_Reshape_1 (TensorFl [(None, 128)]        0           spectral_normalization_6[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_mul (TensorFlowOpLa [(None, 128)]        0           global_sum_pooling2d[0][0]       \n",
      "                                                                 tf_op_layer_Reshape_1[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "spectral_normalization_7 (Spect (None, 1)            130         global_sum_pooling2d[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_Sum (TensorFlowOpLa [(None, 1)]          0           tf_op_layer_mul[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "tf_op_layer_add_4 (TensorFlowOp [(None, 1)]          0           spectral_normalization_7[0][0]   \n",
      "                                                                 tf_op_layer_Sum[0][0]            \n",
      "==================================================================================================\n",
      "Total params: 269,570\n",
      "Trainable params: 268,865\n",
      "Non-trainable params: 705\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "discriminator.summary()\n",
    "plot_model(discriminator, show_shapes=True, dpi=64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = tf.data.Dataset.from_tensor_slices((train_data, oh_labels, labels)).repeat().shuffle(len(train_data)).batch(BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "D_optimizer = Adam(learning_rate=LR, beta_1=BETA_1, beta_2=BETA_2)\n",
    "G_optimizer = Adam(learning_rate=LR, beta_1=BETA_1, beta_2=BETA_2)\n",
    "\n",
    "def learning_rate_decay(current_lr, decay_factor=DECAY_FACTOR):\n",
    "    new_lr = max(current_lr / decay_factor, MIN_LR)\n",
    "    return new_lr\n",
    "\n",
    "def set_learning_rate(new_lr):\n",
    "    '''\n",
    "        Set new learning rate to optimizers\n",
    "    '''\n",
    "    \n",
    "    K.set_value(D_optimizer.lr, new_lr)\n",
    "    K.set_value(G_optimizer.lr, new_lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# DC LOSS\n",
    "\n",
    "# cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True)\n",
    "# def discriminator_loss(real_output, fake_output):\n",
    "#     real_loss = cross_entropy(tf.ones_like(real_output), real_output)\n",
    "#     fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)\n",
    "#     total_loss = real_loss + fake_loss\n",
    "#     return total_loss\n",
    "\n",
    "# def generator_loss(fake_output):\n",
    "#     return cross_entropy(tf.ones_like(fake_output), fake_output)\n",
    "\n",
    "# def discriminator_loss(real_output, fake_output):\n",
    "#     real_loss = tf.reduce_mean(tf.math.softplus(-real_output))\n",
    "#     fake_loss = tf.reduce_mean(tf.math.softplus(fake_output))\n",
    "#     total_loss = real_loss + fake_loss\n",
    "#     return total_loss\n",
    "\n",
    "# def generator_loss(fake_output):\n",
    "#     return tf.reduce_mean(tf.math.softplus(-fake_output))\n",
    "\n",
    "#Hinge loss\n",
    "def discriminator_loss(real_output, fake_output):\n",
    "    real_loss = tf.reduce_mean(tf.nn.relu(tf.ones_like(real_output) - real_output))\n",
    "    fake_loss = tf.reduce_mean(tf.nn.relu(tf.ones_like(fake_output) + fake_output))\n",
    "    total_loss = real_loss + fake_loss\n",
    "    return total_loss\n",
    "\n",
    "def generator_loss(fake_output):\n",
    "    return -tf.reduce_mean(fake_output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint_path = os.path.join(\"checkpoints\", \"tensorflow\")\n",
    "\n",
    "ckpt = tf.train.Checkpoint(generator=generator,\n",
    "                           discriminator=discriminator,\n",
    "                           G_optimizer=G_optimizer,\n",
    "                           D_optimizer=D_optimizer)\n",
    "\n",
    "ckpt_manager = tf.train.CheckpointManager(ckpt, checkpoint_path, max_to_keep=5)\n",
    "\n",
    "# if a checkpoint exists, restore the latest checkpoint.\n",
    "if ckpt_manager.latest_checkpoint:\n",
    "    ckpt.restore(ckpt_manager.latest_checkpoint)\n",
    "    latest_epoch = int(ckpt_manager.latest_checkpoint.split('-')[1])\n",
    "    CURRENT_EPOCH = latest_epoch * SAVE_EVERY_N_EPOCH\n",
    "    print ('Latest checkpoint of epoch {} restored!!'.format(CURRENT_EPOCH))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_and_save_images(model, epoch, test_input, save=True, is_flatten=True):\n",
    "    # Test input is a list include noise and label\n",
    "    predictions = model.predict(test_input)\n",
    "    if is_flatten:\n",
    "        predictions = predictions.reshape(-1, 28, 28, 1).astype('float32')\n",
    "    fig = plt.figure(figsize=(3,3))\n",
    "    for i in range(predictions.shape[0]):\n",
    "        plt.subplot(3, 3, i+1)\n",
    "        plt.imshow(predictions[i, :, :, 0] * 127.5 + 127.5, cmap='gray')\n",
    "        plt.axis('off')\n",
    "    \n",
    "    if save:\n",
    "        plt.savefig(os.path.join(OUTPUT_PATH, 'image_at_epoch_{:04d}.png'.format(epoch)))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x216 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_examples_to_generate = 9\n",
    "\n",
    "# We will reuse this seed overtime (so it's easier)\n",
    "# to visualize progress in the animated GIF)\n",
    "sample_noise = tf.random.normal([num_examples_to_generate, NOISE_DIM])\n",
    "sample_label = np.arange(0, 9).reshape(9, 1)\n",
    "generate_and_save_images(generator, 0, [sample_noise, sample_label], save=False, is_flatten=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def train_d_step(real_image, real_label, batch_size, step):\n",
    "    '''\n",
    "        One training D step\n",
    "        \n",
    "        Reference: https://www.tensorflow.org/tutorials/generative/dcgan\n",
    "    '''\n",
    "    print(\"retrace\")\n",
    "    noise = tf.random.normal([batch_size, NOISE_DIM])\n",
    "    # Open a GradientTape to record the operations run\n",
    "    # during the forward pass, which enables autodifferentiation.\n",
    "    with tf.GradientTape(persistent=True) as tape:\n",
    "        fake_image = generator([noise, real_label], training=True)\n",
    "        \n",
    "        real_pred = discriminator([real_image, real_label], training=True)\n",
    "        fake_pred = discriminator([fake_image, real_label], training=True)\n",
    "        \n",
    "        D_loss = discriminator_loss(real_pred, fake_pred)\n",
    "    # Calculate the gradients for discriminator\n",
    "    D_gradients = tape.gradient(D_loss, \n",
    "                                            discriminator.trainable_variables)\n",
    "\n",
    "    # Apply the gradients to the optimizer\n",
    "    D_optimizer.apply_gradients(zip(D_gradients,\n",
    "                                                discriminator.trainable_variables))\n",
    "    \n",
    "    # Write loss values to tensorboard\n",
    "    if step % 10 == 0:\n",
    "        with file_writer.as_default():\n",
    "            tf.summary.scalar('D_loss', D_loss, step=step)\n",
    "            \n",
    "@tf.function\n",
    "def train_g_step(real_image, real_label, batch_size, step):\n",
    "    '''\n",
    "        One training G step\n",
    "        \n",
    "        Reference: https://www.tensorflow.org/tutorials/generative/dcgan\n",
    "    '''\n",
    "    print(\"retrace\")\n",
    "    noise = tf.random.normal([batch_size, NOISE_DIM])\n",
    "    # Open a GradientTape to record the operations run\n",
    "    # during the forward pass, which enables autodifferentiation.\n",
    "    with tf.GradientTape(persistent=True) as tape:\n",
    "        fake_image = generator([noise, real_label], training=True)\n",
    "        \n",
    "        fake_pred = discriminator([fake_image, real_label], training=True)\n",
    "        \n",
    "        G_loss = generator_loss(fake_pred)\n",
    "    # Calculate the gradients for discriminator\n",
    "    G_gradients = tape.gradient(G_loss, \n",
    "                                        generator.trainable_variables)\n",
    "\n",
    "    # Apply the gradients to the optimizer\n",
    "    G_optimizer.apply_gradients(zip(G_gradients, \n",
    "                                            generator.trainable_variables))\n",
    "    \n",
    "    # Write loss values to tensorboard\n",
    "    if step % 10 == 0:\n",
    "        with file_writer.as_default():\n",
    "            tf.summary.scalar('G_loss', G_loss, step=step)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current iterations: 100001\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x216 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving checkpoint for epoch 100001 at checkpoints\\tensorflow\\ckpt-20\n",
      "Time taken for 5000 iters is 161.6469268798828 sec\n",
      "\n",
      "Saving checkpoint for epoch 100001 at checkpoints\\tensorflow\\ckpt-21\n"
     ]
    }
   ],
   "source": [
    "current_learning_rate = LR\n",
    "trace = True\n",
    "d_training_count = 0\n",
    "start = time.time()\n",
    "\n",
    "for step, (image, oh_label, label) in enumerate(train_data):\n",
    "    if step > DECAY_ITERATIONS:\n",
    "        current_learning_rate = learning_rate_decay(current_learning_rate)\n",
    "        set_learning_rate(current_learning_rate)\n",
    "    \n",
    "    batch_size = tf.constant(label.shape[0], dtype=tf.int32)\n",
    "    train_d_step(image, label, batch_size, step=tf.constant(step, dtype=tf.int64))\n",
    "    d_training_count += 1\n",
    "    if d_training_count >= D_G_TRAINING_RATIO:\n",
    "        train_g_step(image, label, batch_size, step=tf.constant(step, dtype=tf.int64))\n",
    "        d_training_count = 0\n",
    "\n",
    "    if step % 10 == 0:\n",
    "        print ('.', end='')\n",
    "    \n",
    "    if step % (SAVE_EVERY_N_ITER / 2) == 0 and step != 0:\n",
    "        # Clear jupyter notebook cell output\n",
    "        clear_output(wait=True)\n",
    "        # Using a consistent image (sample_X) so that the progress of the model is clearly visible.\n",
    "        print ('Current iterations: {}'.format(CURRENT_ITERATION))\n",
    "        generate_and_save_images(generator, CURRENT_ITERATION, [sample_noise, sample_label], save=True)\n",
    "    \n",
    "    if step % SAVE_EVERY_N_ITER == 0 and step != 0:\n",
    "        ckpt_save_path = ckpt_manager.save()\n",
    "        print ('Saving checkpoint for epoch {} at {}'.format(CURRENT_ITERATION,\n",
    "                                                             ckpt_save_path))\n",
    "    \n",
    "        print ('Time taken for {} iters is {} sec\\n'.format(SAVE_EVERY_N_ITER,\n",
    "                                                          time.time()-start))\n",
    "        start = time.time()\n",
    "        \n",
    "    if CURRENT_ITERATION > ITERATIONS:\n",
    "        # Save at final\n",
    "        ckpt_save_path = ckpt_manager.save()\n",
    "        print ('Saving checkpoint for epoch {} at {}'.format(CURRENT_ITERATION,\n",
    "                                                                ckpt_save_path))\n",
    "        break\n",
    "    CURRENT_ITERATION += 1"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}