Training_Gans_SpectralNormalization_MPS60.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    }
   ],
   "source": [
    "from Model.GeoGans import GAN2D_MPS, WGAN2D_MPS, AlphaGAN_MPS, CycleGAN_MPS\n",
    "from Model.Utils import PlotHistory, Save_Model,PlotDataAE\n",
    "from Model.Utils import load_numpy, convert_to_tfrecords, MPS_Generator, model_plot\n",
    "from keras.optimizers import RMSprop,Adam\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "import tensorflow as tf\n",
    "import os\n",
    "import timeit\n",
    "%matplotlib notebook"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load DataSet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "load_tfrecords = True \n",
    "create_tfrecords = False\n",
    "\n",
    "path_tfr = 'DataSet/MPS60'\n",
    "path_npy = '/share/GeoFacies/ModelosDeFacies/ModelMPS60/DataSet/DataSet.npy'\n",
    "batch_size = 32"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "if load_tfrecords:    \n",
    "    if create_tfrecords: \n",
    "        x_train, x_test = load_numpy(path_npy,random_state=0,split_data=0.30)\n",
    "        convert_to_tfrecords(path_tfr,x_train,x_test)\n",
    "        gen_train = MPS_Generator(os.path.join(path_tfr, 'train.tfrecords'), 14000)\n",
    "        \n",
    "    else:\n",
    "        try:\n",
    "            gen_train = MPS_Generator(os.path.join(path_tfr, 'train.tfrecords'), 14000)\n",
    "        except:\n",
    "            print(\"Data not found. Change 'create_tfrecords' to True\")\n",
    "    x_train = gen_train.get_numpy_batch().astype('float32')\n",
    "else:\n",
    "    x_train, _ = load_numpy(path_npy,random_state=0,split_data=0.30)\n",
    "x_train = x_train*2 -1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training GAN models "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "_, n_rows, n_cols, n_chns = x_train.shape\n",
    "path_weights = os.path.join(path_tfr, 'w60_GAN.hdf5')\n",
    "num_epochs = 25"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. GAN2D Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gan_g_input (InputLayer)     (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "gan_g_dense (Dense)          (None, 16384)             8208384   \n",
      "_________________________________________________________________\n",
      "gan_g_reshape (Reshape)      (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_g_bn_dense (BatchNormali (None, 8, 8, 256)         1024      \n",
      "_________________________________________________________________\n",
      "gan_g_relu (Activation)      (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_g_1_deconv2D (Conv2DTran (None, 16, 16, 128)       524416    \n",
      "_________________________________________________________________\n",
      "gan_g_1_bn (BatchNormalizati (None, 16, 16, 128)       512       \n",
      "_________________________________________________________________\n",
      "gan_g_1_relu (Activation)    (None, 16, 16, 128)       0         \n",
      "_________________________________________________________________\n",
      "gan_g_bilinear (BilinearUpsa (None, 15, 15, 128)       0         \n",
      "_________________________________________________________________\n",
      "gan_g_2_deconv2D (Conv2DTran (None, 30, 30, 64)        131136    \n",
      "_________________________________________________________________\n",
      "gan_g_2_bn (BatchNormalizati (None, 30, 30, 64)        256       \n",
      "_________________________________________________________________\n",
      "gan_g_2_relu (Activation)    (None, 30, 30, 64)        0         \n",
      "_________________________________________________________________\n",
      "gan_g_3_deconv2D (Conv2DTran (None, 60, 60, 32)        32800     \n",
      "_________________________________________________________________\n",
      "gan_g_3_bn (BatchNormalizati (None, 60, 60, 32)        128       \n",
      "_________________________________________________________________\n",
      "gan_g_3_relu (Activation)    (None, 60, 60, 32)        0         \n",
      "_________________________________________________________________\n",
      "gan_g_4_deconv2D (Conv2DTran (None, 60, 60, 2)         1026      \n",
      "_________________________________________________________________\n",
      "gan_g_4_tanh (Activation)    (None, 60, 60, 2)         0         \n",
      "=================================================================\n",
      "Total params: 8,899,682\n",
      "Trainable params: 8,898,722\n",
      "Non-trainable params: 960\n",
      "_________________________________________________________________\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gan_d_input (InputLayer)     (None, 60, 60, 2)         0         \n",
      "_________________________________________________________________\n",
      "gan_d_1_conv2D_SN (ConvSN2D) (None, 30, 30, 32)        1088      \n",
      "_________________________________________________________________\n",
      "gan_d_1_leaky_relu (LeakyReL (None, 30, 30, 32)        0         \n",
      "_________________________________________________________________\n",
      "gan_d_2_conv2D_SN (ConvSN2D) (None, 15, 15, 64)        32896     \n",
      "_________________________________________________________________\n",
      "gan_d_2_leaky_relu (LeakyReL (None, 15, 15, 64)        0         \n",
      "_________________________________________________________________\n",
      "gan_d_3_conv2D_SN (ConvSN2D) (None, 8, 8, 128)         131328    \n",
      "_________________________________________________________________\n",
      "gan_d_3_leaky_relu (LeakyReL (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "gan_d_4_conv2D_SN (ConvSN2D) (None, 8, 8, 256)         524800    \n",
      "_________________________________________________________________\n",
      "gan_d_4_leaky_relu (LeakyReL (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_d_flatten (Flatten)      (None, 16384)             0         \n",
      "_________________________________________________________________\n",
      "gan_d_dense_SN (DenseSN)     (None, 1)                 16386     \n",
      "_________________________________________________________________\n",
      "gan_d_sigmoid (Activation)   (None, 1)                 0         \n",
      "=================================================================\n",
      "Total params: 706,498\n",
      "Trainable params: 706,017\n",
      "Non-trainable params: 481\n",
      "_________________________________________________________________\n",
      "WARNING:tensorflow:From /work/Smith/.local/lib/python3.5/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gan_e_input (InputLayer)     (None, 60, 60, 2)         0         \n",
      "_________________________________________________________________\n",
      "gan_e_1_conv2D (Conv2D)      (None, 30, 30, 32)        1056      \n",
      "_________________________________________________________________\n",
      "gan_e_1_leaky_relu (LeakyReL (None, 30, 30, 32)        0         \n",
      "_________________________________________________________________\n",
      "gan_e_2_conv2D (Conv2D)      (None, 15, 15, 64)        32832     \n",
      "_________________________________________________________________\n",
      "gan_e_2_bn (BatchNormalizati (None, 15, 15, 64)        256       \n",
      "_________________________________________________________________\n",
      "gan_e_2_leaky_relu (LeakyReL (None, 15, 15, 64)        0         \n",
      "_________________________________________________________________\n",
      "gan_e_3_conv2D (Conv2D)      (None, 8, 8, 128)         131200    \n",
      "_________________________________________________________________\n",
      "gan_e_3_bn (BatchNormalizati (None, 8, 8, 128)         512       \n",
      "_________________________________________________________________\n",
      "gan_e_3_leaky_relu (LeakyReL (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "gan_e_4_conv2D (Conv2D)      (None, 8, 8, 256)         524544    \n",
      "_________________________________________________________________\n",
      "gan_e_4_bn (BatchNormalizati (None, 8, 8, 256)         1024      \n",
      "_________________________________________________________________\n",
      "gan_e_4_leaky_relu (LeakyReL (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_e_flatten (Flatten)      (None, 16384)             0         \n",
      "_________________________________________________________________\n",
      "gan_e_dropout (Dropout)      (None, 16384)             0         \n",
      "_________________________________________________________________\n",
      "gan_e_dense (Dense)          (None, 500)               8192500   \n",
      "_________________________________________________________________\n",
      "gan_e_bn_out (BatchNormaliza (None, 500)               1000      \n",
      "=================================================================\n",
      "Total params: 8,884,924\n",
      "Trainable params: 8,883,028\n",
      "Non-trainable params: 1,896\n",
      "_________________________________________________________________\n",
      "Time:  2.350378781557083e-05\n"
     ]
    }
   ],
   "source": [
    "model = GAN2D_MPS(input_shape = (n_rows, n_cols, n_chns),SpectralNormalization=True) # saving_path = path_tfr + '/', summary = True\n",
    "\n",
    "start = timeit.default_timer()\n",
    "#model.train(x_train , epochs=num_epochs, num_plots = 10)\n",
    "stop = timeit.default_timer()\n",
    "print('Time: ', stop - start) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 24 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_plot(x_train[:20], model, num_plots = 12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x144 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_random = np.random.normal(0, 1, size=(20, 500))\n",
    "model_plot(x_random, model, num_plots = 12, mode = \"gen\", figsize = (60,60), type_model = \"CycleGAN\" )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. WGAN2D_MPS Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gan_g_input (InputLayer)     (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "gan_g_dense (Dense)          (None, 16384)             8208384   \n",
      "_________________________________________________________________\n",
      "gan_g_reshape (Reshape)      (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_g_bn_dense (BatchNormali (None, 8, 8, 256)         1024      \n",
      "_________________________________________________________________\n",
      "gan_g_relu (Activation)      (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_g_1_deconv2D (Conv2DTran (None, 16, 16, 128)       524416    \n",
      "_________________________________________________________________\n",
      "gan_g_1_bn (BatchNormalizati (None, 16, 16, 128)       512       \n",
      "_________________________________________________________________\n",
      "gan_g_1_relu (Activation)    (None, 16, 16, 128)       0         \n",
      "_________________________________________________________________\n",
      "gan_g_bilinear (BilinearUpsa (None, 15, 15, 128)       0         \n",
      "_________________________________________________________________\n",
      "gan_g_2_deconv2D (Conv2DTran (None, 30, 30, 64)        131136    \n",
      "_________________________________________________________________\n",
      "gan_g_2_bn (BatchNormalizati (None, 30, 30, 64)        256       \n",
      "_________________________________________________________________\n",
      "gan_g_2_relu (Activation)    (None, 30, 30, 64)        0         \n",
      "_________________________________________________________________\n",
      "gan_g_3_deconv2D (Conv2DTran (None, 60, 60, 32)        32800     \n",
      "_________________________________________________________________\n",
      "gan_g_3_bn (BatchNormalizati (None, 60, 60, 32)        128       \n",
      "_________________________________________________________________\n",
      "gan_g_3_relu (Activation)    (None, 60, 60, 32)        0         \n",
      "_________________________________________________________________\n",
      "gan_g_4_deconv2D (Conv2DTran (None, 60, 60, 2)         1026      \n",
      "_________________________________________________________________\n",
      "gan_g_4_tanh (Activation)    (None, 60, 60, 2)         0         \n",
      "=================================================================\n",
      "Total params: 8,899,682\n",
      "Trainable params: 8,898,722\n",
      "Non-trainable params: 960\n",
      "_________________________________________________________________\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gan_d_input (InputLayer)     (None, 60, 60, 2)         0         \n",
      "_________________________________________________________________\n",
      "gan_d_1_conv2D_SN (ConvSN2D) (None, 30, 30, 32)        1664      \n",
      "_________________________________________________________________\n",
      "gan_d_1_leaky_relu (LeakyReL (None, 30, 30, 32)        0         \n",
      "_________________________________________________________________\n",
      "gan_d_2_conv2D_SN (ConvSN2D) (None, 15, 15, 64)        51328     \n",
      "_________________________________________________________________\n",
      "gan_d_2_leaky_relu (LeakyReL (None, 15, 15, 64)        0         \n",
      "_________________________________________________________________\n",
      "gan_d_3_conv2D_SN (ConvSN2D) (None, 8, 8, 128)         205056    \n",
      "_________________________________________________________________\n",
      "gan_d_3_leaky_relu (LeakyReL (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "gan_d_4_conv2D_SN (ConvSN2D) (None, 8, 8, 256)         819712    \n",
      "_________________________________________________________________\n",
      "gan_d_4_leaky_relu (LeakyReL (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_d_flatten (Flatten)      (None, 16384)             0         \n",
      "_________________________________________________________________\n",
      "gan_d_dense_SN (DenseSN)     (None, 1)                 16386     \n",
      "_________________________________________________________________\n",
      "gan_d_sigmoid (Activation)   (None, 1)                 0         \n",
      "=================================================================\n",
      "Total params: 1,094,146\n",
      "Trainable params: 1,093,665\n",
      "Non-trainable params: 481\n",
      "_________________________________________________________________\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gan_e_input (InputLayer)     (None, 60, 60, 2)         0         \n",
      "_________________________________________________________________\n",
      "gan_e_1_conv2D (Conv2D)      (None, 30, 30, 32)        1056      \n",
      "_________________________________________________________________\n",
      "gan_e_1_leaky_relu (LeakyReL (None, 30, 30, 32)        0         \n",
      "_________________________________________________________________\n",
      "gan_e_2_conv2D (Conv2D)      (None, 15, 15, 64)        32832     \n",
      "_________________________________________________________________\n",
      "gan_e_2_bn (BatchNormalizati (None, 15, 15, 64)        256       \n",
      "_________________________________________________________________\n",
      "gan_e_2_leaky_relu (LeakyReL (None, 15, 15, 64)        0         \n",
      "_________________________________________________________________\n",
      "gan_e_3_conv2D (Conv2D)      (None, 8, 8, 128)         131200    \n",
      "_________________________________________________________________\n",
      "gan_e_3_bn (BatchNormalizati (None, 8, 8, 128)         512       \n",
      "_________________________________________________________________\n",
      "gan_e_3_leaky_relu (LeakyReL (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "gan_e_4_conv2D (Conv2D)      (None, 8, 8, 256)         524544    \n",
      "_________________________________________________________________\n",
      "gan_e_4_bn (BatchNormalizati (None, 8, 8, 256)         1024      \n",
      "_________________________________________________________________\n",
      "gan_e_4_leaky_relu (LeakyReL (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "gan_e_flatten (Flatten)      (None, 16384)             0         \n",
      "_________________________________________________________________\n",
      "gan_e_dropout (Dropout)      (None, 16384)             0         \n",
      "_________________________________________________________________\n",
      "gan_e_dense (Dense)          (None, 500)               8192500   \n",
      "_________________________________________________________________\n",
      "gan_e_bn_out (BatchNormaliza (None, 500)               1000      \n",
      "=================================================================\n",
      "Total params: 8,884,924\n",
      "Trainable params: 8,883,028\n",
      "Non-trainable params: 1,896\n",
      "_________________________________________________________________\n",
      "WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Initializing GANs models variables ...\n",
      "Starting GANs training ...\n",
      "Saving check point ...\n",
      "Iteration:  5000   D loss:  -131.42867   G loss:  -2148.7412   Accuracy:  0.97125\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  10000   D loss:  -35.174614   G loss:  -2246.3396   Accuracy:  0.9653212\n",
      "Saving check point ...\n",
      "Iteration:  15000   D loss:  -29.835487   G loss:  -2186.964   Accuracy:  0.9645747\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  20000   D loss:  -29.722128   G loss:  -2178.6475   Accuracy:  0.9678863\n",
      "Iteration:  25000   D loss:  -30.107435   G loss:  -2078.7712   Accuracy:  0.9588455\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration:  30000   D loss:  -30.329319   G loss:  -2000.7384   Accuracy:  0.95338976\n",
      "Iteration:  35000   D loss:  -30.406872   G loss:  -1938.1357   Accuracy:  0.9604123\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABqIAAAGRCAYAAAAO3qLvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAGn9JREFUeJzt3c+rLctVB/BVsiMSREUyiILxITjyHxBBUFRQhw4EQUFBnAmC0YEotQsEUTJz4DBCRjpwIojEH+BInEUyFXxiSBQyUEmEoFAOzn3em5fd55w+p1bX7u7PZ3TZ99zutXt31ao+31dvl957AAAAAAAAwGjfNLsAAAAAAAAAjkkQBQAAAAAAQApBFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFGcQinlF0spny+l/Hcp5d9KKX9USvmOZ/7b90spPz6wlqHHA+Bc9DQAjkA/A+Ao9DR4miCKwyul/HpE/H5E/EZEfHtE/GBEfG9E/FUp5Ztn1gYAa+hpAByBfgbAUehp8DyCKA6tlPJtEdEi4ld773/Ze/+f3vv7EfGzEfFeRPx8KeWPSym/+86/+ZFSyhfe/PkzEfGJiPjzUspXSim/WUp5r5TSSym/Ukr5YinlS6WUT77z71cdL/0iAHAIehoAR6CfAXAUeho832V2AZDshyLiWyLiz959sff+lVLKX0TET0TE15b+ce/9F0opPxwRv9x7/+uIiFLKe2/++kcj4vsj4vsi4m9LKZ/74GfWHA8AnklPA+AI9DMAjkJPg2eyI4qj+1hEfLn3/r83/u5Lb/7+pVrv/au9989HxKcj4udecSwAeIqeBsAR6GcAHIWeBs8kiOLovhwRHyul3Nr9911v/v6l/vWdP/9LRHz3K44FAE/R0wA4Av0MgKPQ0+CZBFEc3d/HwxbYn3n3xVLKt0bET0XE30TEVyPio+/89cc/dIy+cOzveefPn4iIL77580uPBwCP0dMAOAL9DICj0NPgmQRRHFrv/T/j4UsD/7CU8pOllI+8+X+t/mlEfCEiPhMRn4uIny6lfGcp5eMR8WsfOsy/x8P/j/XDfqeU8tFSyg9ExC9FxJ+8ef2lxwOARXoaAEegnwFwFHoaPJ8gisPrvf9BRPxWRHwqIv4rIv4hHra3/ljv/Wvx0BT+MSLej4jPxtuJ/QO/FxG/XUr5j1LKJ995/e8i4p/i4b9u+FTv/bNvXn/p8QDgUXoaAEegnwFwFHoaPE/p3W49WOPNf9nwzxHxkYUvIwSAXdDTADgC/QyAo9DTOCo7ogAAAAAAAEghiAIAAAAAACCF/zUfAAAAAAAAKeyIAgAAAAAAIMVly5OVcrX9it2o0WaXQIIW9ebrmZ/30jnXGlXjtfcy5EAnd6aeNms+HDV2gJe5NfbXjsul+WNUP9bTXu9aymn62b3R58bw3HZcZxojvV/1swH20NPOdF9zf/TMeWPwTNd+6RnNjigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAghSAKAAAAAACAFIIoAAAAAAAAUpTe+3YnK9ftTgYJarTZJRxeizq7hMNYul+vvZeNSzmkM/W07LnPuAeeoqfluZZymn52JrN6q+clZtrzmrL3q342wNpntKU5a8/3Esy053VA9rjf87VZa+kZzY4oAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSCKAAAAAAAAFKU3vt2JyvX7U4Gr1SjzS7h0FrU2SWcVu/XMruGI1jb09bOKUtjZMTctHb8LZ3TOAZm09Neb1Y/G8F6HXiOPaxZ9bMx/N4R7tMR12yZv7PZu2vvN3uaHVEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApLjMLoC5arTZJTypRZ1dAq/g84MHxgIAR6CfAXuz5vce5jgAnmMPv1O/N3ZEAQAAAAAAkEIQBQAAAAAAQApBFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJDiMruAkWq02SWQwOc6T4s6uwQ4jHuay0bVsnQccwcAW8juN/fUu4H7ZD0M8LSlOXHPa60jvqdsdkQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApBBEAQAAAAAAkKL03jc72bWU7U4GPFuLOrsENtT7tcyu4Qj0tGXmFOClarRVP3/tXU97pVKu+tkga+9f7ov1i3t4Jv1sDD0NjkuP2o+lnmZHFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAECKy+wCgJfzhbpwXEvj2xd0AsB9GrU21+vH8Ky0XvY1O9O9bS0PMNZe+vo9zfOzrtl14XU7ogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIUXrvm53sWsp2J4M71qLOLoET6/1aZtdwBKVcp/S0Gm3GaYcw9wEfGDWXXXvX015pVj8j117WC9YGbGEP40E/G2Pp947mGoDtLP3e0Y4oAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSCKAAAAAAAAFJcZhcAADxfi/oNr9VoEyoBeJr5Cea4tV4AODpzH8D9siMKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSX2QWwjRb15us12qt+9myWrg3ATGvnJvM5AABnsWatbJ0MADnsiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAgxWV2AYzVoqb9/NpjL6nRhhwHgH1Zmv9H9RdgHus7AI5g7bpU/wOA57EjCgAAAAAAgBSCKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUl9kFjNSi3ny9Rks7Nuu5lgBzZfZL4NjMEwAAAKxlRxQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKS6zC9hCizq7BABgQY025Dj6PTwYNaYAgMctrT/1YgD4enZEAQAAAAAAkEIQBQAAAAAAQApBFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJDisuXJWtQtTwcAAAAAm1r7+69rThkAcDfsiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAgxWV2AQAAI9Ro3/BaizqhEhjr1r0NAAAAe2FHFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApLrMLAAC2VaPNLmEza99ri5pUCTztTGMTAACA87AjCgAAAAAAgBSCKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUl9kFAADbalFvvl6jbVwJnJOxtmxpflpyzSnjVNbej2s/IwDeWp5zr1uWAQCbsyMKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSX2QUAAPehRb35eo22cSXzrH2vS9eMcznTGFnLGDke8yTAW9YAAPA8dkQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApLjMLgAAuG9rv2j+TF/avPRe114z7suZ7uG13NusZZ4EjsDaAABex44oAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSCKAAAAAAAAFJcZhcAABxLi3rz9Rpt40rmOdN7zeZ+mmfp2sMIa8awexEYzToCAL7euN54vfmqHVEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApLjMLgAAOIcWddXP12hJlbAn7oNx1o5BuBej5gFjAACAs5v1jG1HFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApLrMLAAC4pUVd9fM1WlIlsC9rxw6cxVKfMGbg2KwRAdibI/YuO6IAAAAAAABIIYgCAAAAAAAghSAKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASHGZXQAAwAgt6je8VqNNqATGunVvA+Os7RXGJNwn6z4A9uZMvcuOKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSXGYXAACQpUVd9fM1WlIl8LS19yuMsHTfmQ/hfIx7AHgePXM9O6IAAAAAAABIIYgCAAAAAAAghSAKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASHGZXQAAz1Oj3Xy9Rd24EjiupfG0NP7gJczb7MGo+/SI82f2ezJHMMoRxx8AvIbeOI8dUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkuMwuAODoarQ7O/41oww4tBb15uvZ4/tMlq7xWjM+k1G1wxGZP2EbxtR9MfcBzGW+vT92RAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQ4jK7AIAjqdFmlwDwLC3q7BJSHPV9wdHcGqvWUQ/WXAdz3nEZD/uwdgwu/fx1QC0AR6EHHpMdUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkuMwuAABgr1rU2SUAcGI12s3X9af9WPoMuS/GFMB4euC52BEFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApBBEAQAAAAAAkEIQBQAAAAAAQIrL7AIA9qhGm10CAMChtKhDjmOdxhbcZ8c0ah4C4C09875k97rrwut2RAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQ4jK7AAAAABilRR1ynBptyHFmWFv7qGt2T/b8+THWEe9vgJn02Puylz5nRxQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABAisvsAgDugS9aBADgXWu/+HnP68k9185+7OXL1AHujT59XGfqjXZEAQAAAAAAkEIQBQAAAAAAQApBFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJDiMrsAgNeo0WaXAMDJZfeiFjX1+MAYa8eqdSx7oQ8BvKV/8xJ6qR1RAAAAAAAAJBFEAQAAAAAAkEIQBQAAAAAAQApBFAAAAAAAACkEUQAAAAAAAKS4zC4AOIYabXYJADCEnpZv6Rq3qBtXAvOsud/NSzzF/AmMtLbv7GEO0kuJ2Me9elR2RAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQ4jK7AACep0UdcpzrkKMA7EuNNruEF9tz7Wutf6/XjDLYwIj7etTaCPbEfQ8c1ZnWvOTTL++PHVEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApLjMLgB4uRrt5ust6qqfB4DZ9Cg4JmN7DOv7/Vj6rABeYi/z/F7qZB/00mOyIwoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAghSAKAAAAAACAFJfZBQBv1Wh3dRxytaizSwB4NT0HjumexrY107Kla3NPn99euM9g/5bmvhnj2zwMT9N7z8WOKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSXGYXAEdWo80ugY21qLNLAHgWPQqIMBdwPtbrcD5LvW5pPtAb4XF6KS9hRxQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKS6zC4DZarTZJbBDLersEgAADsX6apzsa3lPz1DuG+Cl7mku24u1c65rPIZexxHYEQUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABAisvsAjifGm12CZxYizq7BIA0eiwAW7CmBji2UfP8rePc2zOLngbbsCMKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSXLU9Wo918vUVNPf6ZjLqWI/g8uEej7st7GmvAfuiN7N1S/3Nvs8SaCQDG23N/3XPtwMvZEQUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABAisvsAiIiarSbr7eoq36e9ddyxLFhphH3NgA8h56zLPvaWIcCAPCuzN+BAuPZEQUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABAisvsAh5To80u4TBcSx7Tor76GO4x4IzMfcc0oi8y1tJnct22DB5h3ABwr9b2qBlr/DP1Uc9QcE52RAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkuMwuAHjaHr60cg81Aufji3DPRz+CdYwZAI5CT7s/I57HZn2uS7W7z+Bl7IgCAAAAAAAghSAKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIMVldgFwRi3q7BIA7l6NNrsEnkFPA2BPRq0v9D84rjON7z08c91bjfdWzwxnGiOMY0cUAAAAAAAAKQRRAAAAAAAApBBEAQAAAAAAkEIQBQAAAAAAQApBFAAAAAAAACkuswuAPWlRZ5cAAEPoaQDwcjXakOPox8AWRs1ZEHGu+0mfHseOKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSXGYXAKO1qLNLADi1Gm12CYen1wEAAB/mWQzGWhpTS8/ka8fgiOOs/f3AqBrXsiMKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSl977dycr15slqtM1q2EqLmnr8Gdcs+z0B2+j9WmbXcATXUm72tBlz5RH76NnosfAyetrrLT2jwVHtZd1kbXAu+tkYR+xpe5mzRhg1753pmsE9uvZ+s6fZEQUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABAitJ73+5k5brdyQC4qfdrmV3DEYzqaTXaiMMwSYs6uwQ4NT3t9TyjwYO9r8msSfZNPxtjzz1t73PQGrPmqzNdY5jp2vvNnmZHFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAEAKQRQAAAAAAAApLrMLAIA9qtFmlwAAABGxbm3aoiZWAjzmbM+R9zTfLNVyts8EZrEjCgAAAAAAgBSCKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUl9kFAADcixZ1dgkAwARLa4AabeNK8q19T9ZH8LgjzhNLjjgfjHpPZ7oP4CXsiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAgxWV2AQAAa7Sos0vgDtRoq37efQPAY9b2lTMZdW30Yrg/xuU4S9dSf4EHdkQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApBBEAQAAAAAAkOIyuwDgrRpt1c+3qEmVAMxnjmPvlvq6exuAM1r7vLvWUn9dc961PVqv37fse/KeuCfnce3HONN4PSo7ogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIUXrv252sXLc7GbxSjTa7hFNqUWeXcHi9X8vsGo7gWsrpe5rxykxn6tNLY23pGpxpbOppr+cZDR6cqa9wf66962cDeEZbdqb1ITxGv8+31NPsiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAgxWV2ATBbjTa7hFNqUWeXAHyIccmRrbm/l9YGS8cYtZZYOwaNWYD1sp//1s7NnkeBl7AOhJfZy9jJXB/MWqvYEQUAAAAAAEAKQRQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQovTetztZuW53MvgQXwK7D3v50sA96/1aZtdwBHoa8IFRaww9cD097fX0M47K3OwZeE+uvetnA1xLOVxP2/McBJzT0jOaHVEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKQRRAAAAAAAApLjMLgBGq9Fml8ArLH1+LerGlcDj3KsAwNF5ttr3mm9Wje4bWG8PcwrAa9gRBQAAAAAAQApBFAAAAAAAACkEUQAAAAAAAKQQRAEAAAAAAJBCEAUAAAAAAECKy+wCAGCPWtTZJQAD1GizS/h/o2pZOz8tndc8B7kyx949zW17Zy5cb1QfgiMypwBnZUcUAAAAAAAAKQRRAAAAAAAApBBEAQAAAAAAkEIQBQAAAAAAQApBFAAAAAAAACkuswuAp9Ros0vgDizdBy3qxpUAcCQj+oi1CuQ62zrw1vtdeq/mH45g7Vh237MHR+1RAC9lRxQAAAAAAAApBFEAAAAAAACkEEQBAAAAAACQQhAFAAAAAABACkEUAAAAAAAAKS6zCwB4jRrt5ust6saVAHAPlvrCmYy6BiOOM6ofL9dyHXJ89ulM4/1M7xWestRb1jwbGlMAsC07ogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIUXrv252sXLc7GXerRptdAifWos4uYbrer2V2DUegp8E2rBt4zLV3Pe2VsvuZMcxeeE7gJdbOcUv3mWe0Ma6lTHlGM38AvLXU0+yIAgAAAAAAIIUgCgAAAAAAgBSCKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACBF6b3PrgEAAAAAAIADsiMKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSCKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABIIYgCAAAAAAAghSAKAAAAAACAFIIoAAAAAAAAUgiiAAAAAAAASCGIAgAAAAAAIIUgCgAAAAAAgBSCKAAAAAAAAFIIogAAAAAAAEghiAIAAAAAACCFIAoAAAAAAIAUgigAAAAAAABSCKIAAAAAAABI8X8ifEtOHeuOTgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration:  40000   D loss:  -30.407274   G loss:  -1878.5724   Accuracy:  0.9476476\n",
      "Iteration:  45000   D loss:  -30.358696   G loss:  -1824.0472   Accuracy:  0.9531337\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration:  50000   D loss:  -30.067087   G loss:  -1765.1688   Accuracy:  0.9519228\n",
      "Iteration:  55000   D loss:  -29.886028   G loss:  -1728.6654   Accuracy:  0.9429688\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  60000   D loss:  -29.705212   G loss:  -1689.9552   Accuracy:  0.9458247\n",
      "Saving check point ...\n",
      "Iteration:  65000   D loss:  -29.531754   G loss:  -1660.8105   Accuracy:  0.9515886\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  70000   D loss:  -29.25532   G loss:  -1626.6918   Accuracy:  0.9310026\n",
      "Saving check point ...\n",
      "Iteration:  75000   D loss:  -28.943693   G loss:  -1612.4482   Accuracy:  0.9506033\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  80000   D loss:  -28.617422   G loss:  -1584.2146   Accuracy:  0.94614583\n",
      "Saving check point ...\n",
      "Iteration:  85000   D loss:  -28.40155   G loss:  -1559.046   Accuracy:  0.95110244\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  90000   D loss:  -27.967375   G loss:  -1535.5844   Accuracy:  0.9492665\n",
      "Saving check point ...\n",
      "Iteration:  95000   D loss:  -27.80508   G loss:  -1508.4633   Accuracy:  0.95006514\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  100000   D loss:  -27.524132   G loss:  -1496.3041   Accuracy:  0.9446962\n",
      "Saving check point ...\n",
      "Iteration:  105000   D loss:  -27.181725   G loss:  -1484.9047   Accuracy:  0.9465625\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  110000   D loss:  -26.920465   G loss:  -1464.3077   Accuracy:  0.9407205\n",
      "Saving check point ...\n",
      "Iteration:  115000   D loss:  -26.686037   G loss:  -1465.0731   Accuracy:  0.95737416\n",
      "Random realizations ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving check point ...\n",
      "Iteration:  120000   D loss:  -26.399324   G loss:  -1449.6901   Accuracy:  0.9431771\n",
      "Saving check point ...\n",
      "Iteration:  125000   D loss:  -26.12427   G loss:  -1446.5878   Accuracy:  0.9491493\n",
      "Restauring best Discriminator and Generator Models ...\n",
      "Initializing encoder variables ...\n",
      "Starting Encoder training ...\n",
      "epoch 1 E loss -> 155.71257 Acc -> 0.8971148\n",
      "epoch 2 E loss -> 136.35083 Acc -> 0.9093854\n",
      "epoch 3 E loss -> 122.660416 Acc -> 0.9180175\n",
      "epoch 4 E loss -> 117.24936 Acc -> 0.9215322\n",
      "epoch 5 E loss -> 111.97266 Acc -> 0.9248681\n",
      "epoch 6 E loss -> 107.621 Acc -> 0.92764026\n",
      "epoch 7 E loss -> 106.204704 Acc -> 0.92850673\n",
      "epoch 8 E loss -> 103.491325 Acc -> 0.9302766\n",
      "epoch 9 E loss -> 102.218445 Acc -> 0.93108594\n",
      "epoch 10 E loss -> 100.30829 Acc -> 0.93235797\n",
      "epoch 11 E loss -> 98.78247 Acc -> 0.9333366\n",
      "epoch 12 E loss -> 99.49835 Acc -> 0.93286717\n",
      "epoch 13 E loss -> 96.33783 Acc -> 0.93482864\n",
      "epoch 14 E loss -> 93.20775 Acc -> 0.93686014\n",
      "epoch 15 E loss -> 94.07804 Acc -> 0.9362966\n",
      "epoch 16 E loss -> 93.26226 Acc -> 0.936779\n",
      "epoch 17 E loss -> 92.54281 Acc -> 0.9373208\n",
      "epoch 18 E loss -> 90.16353 Acc -> 0.93883675\n",
      "epoch 19 E loss -> 90.212975 Acc -> 0.93880826\n",
      "epoch 20 E loss -> 90.31266 Acc -> 0.9387744\n",
      "epoch 21 E loss -> 88.93329 Acc -> 0.93954194\n",
      "epoch 22 E loss -> 87.945816 Acc -> 0.94017464\n",
      "epoch 23 E loss -> 86.439545 Acc -> 0.94118243\n",
      "epoch 24 E loss -> 86.687454 Acc -> 0.9410149\n",
      "epoch 25 E loss -> 84.79115 Acc -> 0.94218075\n",
      "epoch 26 E loss -> 85.6014 Acc -> 0.9416343\n",
      "epoch 27 E loss -> 84.68807 Acc -> 0.9422698\n",
      "epoch 28 E loss -> 84.775894 Acc -> 0.94217\n",
      "epoch 29 E loss -> 84.8882 Acc -> 0.9421683\n",
      "epoch 30 E loss -> 83.74396 Acc -> 0.94279\n",
      "epoch 31 E loss -> 83.52924 Acc -> 0.942979\n",
      "epoch 32 E loss -> 83.096214 Acc -> 0.94332176\n",
      "epoch 33 E loss -> 81.89619 Acc -> 0.9440831\n",
      "epoch 34 E loss -> 82.78761 Acc -> 0.9435111\n",
      "epoch 35 E loss -> 82.28107 Acc -> 0.94380796\n",
      "epoch 36 E loss -> 82.33475 Acc -> 0.9437655\n",
      "epoch 37 E loss -> 81.36405 Acc -> 0.94437844\n",
      "epoch 38 E loss -> 80.95333 Acc -> 0.9446132\n",
      "epoch 39 E loss -> 80.353806 Acc -> 0.9450477\n",
      "epoch 40 E loss -> 80.20726 Acc -> 0.94505227\n",
      "epoch 41 E loss -> 81.623085 Acc -> 0.9441875\n",
      "epoch 42 E loss -> 79.91017 Acc -> 0.9453332\n",
      "epoch 43 E loss -> 79.85071 Acc -> 0.9453985\n",
      "epoch 44 E loss -> 78.54326 Acc -> 0.94617546\n",
      "epoch 45 E loss -> 80.304085 Acc -> 0.9450658\n",
      "epoch 46 E loss -> 79.475914 Acc -> 0.94558334\n",
      "epoch 47 E loss -> 79.40357 Acc -> 0.9456043\n",
      "epoch 48 E loss -> 78.287346 Acc -> 0.94629824\n",
      "epoch 49 E loss -> 78.330795 Acc -> 0.9463499\n",
      "epoch 50 E loss -> 77.95595 Acc -> 0.94659245\n",
      "epoch 51 E loss -> 78.28298 Acc -> 0.9462936\n",
      "epoch 52 E loss -> 77.243614 Acc -> 0.94699085\n",
      "epoch 53 E loss -> 76.89216 Acc -> 0.9471567\n",
      "epoch 54 E loss -> 76.70407 Acc -> 0.947354\n",
      "epoch 55 E loss -> 76.475746 Acc -> 0.9475442\n",
      "epoch 56 E loss -> 77.69248 Acc -> 0.946664\n",
      "epoch 57 E loss -> 77.05402 Acc -> 0.94716555\n",
      "epoch 58 E loss -> 76.90556 Acc -> 0.9471867\n",
      "epoch 59 E loss -> 76.33288 Acc -> 0.94753075\n",
      "epoch 60 E loss -> 76.239784 Acc -> 0.94763684\n",
      "epoch 61 E loss -> 75.89603 Acc -> 0.9478926\n",
      "epoch 62 E loss -> 76.33771 Acc -> 0.9475946\n",
      "epoch 63 E loss -> 75.80546 Acc -> 0.9478638\n",
      "epoch 64 E loss -> 75.6353 Acc -> 0.94802314\n",
      "epoch 65 E loss -> 75.3804 Acc -> 0.9482739\n",
      "epoch 66 E loss -> 75.36816 Acc -> 0.94812113\n",
      "epoch 67 E loss -> 75.28901 Acc -> 0.948259\n",
      "epoch 68 E loss -> 74.71183 Acc -> 0.94857806\n",
      "epoch 69 E loss -> 74.4256 Acc -> 0.9487619\n",
      "epoch 70 E loss -> 74.48855 Acc -> 0.94871086\n",
      "epoch 71 E loss -> 75.34474 Acc -> 0.9481716\n",
      "epoch 72 E loss -> 74.369965 Acc -> 0.94878316\n",
      "epoch 73 E loss -> 74.485435 Acc -> 0.94869524\n",
      "epoch 74 E loss -> 74.95723 Acc -> 0.948446\n",
      "epoch 75 E loss -> 74.197815 Acc -> 0.9489271\n",
      "epoch 76 E loss -> 74.68636 Acc -> 0.94859916\n",
      "epoch 77 E loss -> 73.742355 Acc -> 0.94924456\n",
      "epoch 78 E loss -> 74.08635 Acc -> 0.94909096\n",
      "epoch 79 E loss -> 73.466385 Acc -> 0.9493925\n",
      "epoch 80 E loss -> 73.89779 Acc -> 0.94912076\n",
      "epoch 81 E loss -> 73.44104 Acc -> 0.94940466\n",
      "epoch 82 E loss -> 73.75692 Acc -> 0.9492139\n",
      "epoch 83 E loss -> 72.84339 Acc -> 0.9497903\n",
      "epoch 84 E loss -> 73.21759 Acc -> 0.9495154\n",
      "epoch 85 E loss -> 72.870834 Acc -> 0.94979465\n",
      "epoch 86 E loss -> 73.641174 Acc -> 0.94932336\n",
      "epoch 87 E loss -> 72.87676 Acc -> 0.9497469\n",
      "epoch 88 E loss -> 73.12994 Acc -> 0.9496524\n",
      "epoch 89 E loss -> 72.5162 Acc -> 0.9500043\n",
      "epoch 90 E loss -> 73.12957 Acc -> 0.9495785\n",
      "epoch 91 E loss -> 73.29838 Acc -> 0.9495448\n",
      "epoch 92 E loss -> 72.238396 Acc -> 0.9501454\n",
      "epoch 93 E loss -> 72.48442 Acc -> 0.95003384\n",
      "epoch 94 E loss -> 72.65998 Acc -> 0.9499534\n",
      "epoch 95 E loss -> 72.186615 Acc -> 0.9501561\n",
      "epoch 96 E loss -> 72.237465 Acc -> 0.9502123\n",
      "epoch 97 E loss -> 72.0302 Acc -> 0.9503516\n",
      "epoch 98 E loss -> 72.06364 Acc -> 0.95029503\n",
      "epoch 99 E loss -> 72.226715 Acc -> 0.9502102\n",
      "epoch 100 E loss -> 71.45082 Acc -> 0.95068294\n",
      "epoch 101 E loss -> 72.1603 Acc -> 0.95020485\n",
      "epoch 102 E loss -> 71.99138 Acc -> 0.95032763\n",
      "epoch 103 E loss -> 71.313354 Acc -> 0.9507441\n",
      "epoch 104 E loss -> 71.82593 Acc -> 0.95044833\n",
      "epoch 105 E loss -> 71.55815 Acc -> 0.95055944\n",
      "epoch 106 E loss -> 71.60668 Acc -> 0.9505525\n",
      "epoch 107 E loss -> 71.69132 Acc -> 0.9505273\n",
      "epoch 108 E loss -> 71.47847 Acc -> 0.9506589\n",
      "epoch 109 E loss -> 72.0152 Acc -> 0.9503163\n",
      "epoch 110 E loss -> 71.41687 Acc -> 0.95070803\n",
      "epoch 111 E loss -> 71.12812 Acc -> 0.9508908\n",
      "epoch 112 E loss -> 71.0113 Acc -> 0.9509401\n",
      "epoch 113 E loss -> 71.45897 Acc -> 0.95065707\n",
      "epoch 114 E loss -> 70.91063 Acc -> 0.9509827\n",
      "epoch 115 E loss -> 70.984604 Acc -> 0.95095634\n",
      "epoch 116 E loss -> 70.75116 Acc -> 0.9511067\n",
      "epoch 117 E loss -> 71.1484 Acc -> 0.9508716\n",
      "epoch 118 E loss -> 71.252235 Acc -> 0.9508082\n",
      "epoch 119 E loss -> 70.63834 Acc -> 0.951214\n",
      "epoch 120 E loss -> 70.99071 Acc -> 0.9509479\n",
      "epoch 121 E loss -> 71.66788 Acc -> 0.9505759\n",
      "epoch 122 E loss -> 71.36508 Acc -> 0.95073146\n",
      "epoch 123 E loss -> 70.317024 Acc -> 0.9513602\n",
      "epoch 124 E loss -> 70.77037 Acc -> 0.9510558\n",
      "epoch 125 E loss -> 70.392944 Acc -> 0.95137566\n",
      "epoch 126 E loss -> 70.53133 Acc -> 0.9512513\n",
      "epoch 127 E loss -> 70.851 Acc -> 0.9510188\n",
      "epoch 128 E loss -> 71.06148 Acc -> 0.9509184\n",
      "epoch 129 E loss -> 71.01642 Acc -> 0.9509569\n",
      "epoch 130 E loss -> 70.748985 Acc -> 0.9511031\n",
      "epoch 131 E loss -> 70.19383 Acc -> 0.9514686\n",
      "epoch 132 E loss -> 70.74327 Acc -> 0.951127\n",
      "epoch 133 E loss -> 70.685455 Acc -> 0.9511599\n",
      "epoch 134 E loss -> 71.08742 Acc -> 0.9508931\n",
      "epoch 135 E loss -> 70.54604 Acc -> 0.9512357\n",
      "epoch 136 E loss -> 70.169075 Acc -> 0.9514855\n",
      "epoch 137 E loss -> 70.15021 Acc -> 0.95150965\n",
      "epoch 138 E loss -> 70.313 Acc -> 0.95140874\n",
      "epoch 139 E loss -> 70.55386 Acc -> 0.9512331\n",
      "epoch 140 E loss -> 70.363144 Acc -> 0.9513469\n",
      "epoch 141 E loss -> 70.01066 Acc -> 0.9515948\n",
      "epoch 142 E loss -> 69.7401 Acc -> 0.9517269\n",
      "epoch 143 E loss -> 70.239845 Acc -> 0.9514634\n",
      "epoch 144 E loss -> 70.22284 Acc -> 0.9514697\n",
      "epoch 145 E loss -> 69.78188 Acc -> 0.95178765\n",
      "epoch 146 E loss -> 69.66696 Acc -> 0.9518062\n",
      "epoch 147 E loss -> 69.90814 Acc -> 0.9516574\n",
      "epoch 148 E loss -> 70.028824 Acc -> 0.95155823\n",
      "epoch 149 E loss -> 69.86313 Acc -> 0.95166415\n",
      "epoch 150 E loss -> 69.31411 Acc -> 0.951983\n",
      "epoch 151 E loss -> 69.57957 Acc -> 0.9519083\n",
      "epoch 152 E loss -> 69.921486 Acc -> 0.9515912\n",
      "epoch 153 E loss -> 69.66795 Acc -> 0.9517848\n",
      "epoch 154 E loss -> 69.6926 Acc -> 0.95182437\n",
      "epoch 155 E loss -> 69.90014 Acc -> 0.95165116\n",
      "epoch 156 E loss -> 69.86895 Acc -> 0.9517103\n",
      "epoch 157 E loss -> 69.67281 Acc -> 0.9518325\n",
      "epoch 158 E loss -> 69.71095 Acc -> 0.9517526\n",
      "epoch 159 E loss -> 69.509285 Acc -> 0.95190364\n",
      "epoch 160 E loss -> 69.535736 Acc -> 0.9519372\n",
      "epoch 161 E loss -> 69.4361 Acc -> 0.9519887\n",
      "epoch 162 E loss -> 70.0 Acc -> 0.9516385\n",
      "epoch 163 E loss -> 69.85108 Acc -> 0.95171195\n",
      "epoch 164 E loss -> 69.53536 Acc -> 0.95190084\n",
      "epoch 165 E loss -> 69.3692 Acc -> 0.9520387\n",
      "epoch 166 E loss -> 70.15869 Acc -> 0.9514902\n",
      "epoch 167 E loss -> 69.451836 Acc -> 0.95196444\n",
      "epoch 168 E loss -> 69.8902 Acc -> 0.95168406\n",
      "epoch 169 E loss -> 69.45077 Acc -> 0.951979\n",
      "epoch 170 E loss -> 69.451096 Acc -> 0.9519069\n",
      "epoch 171 E loss -> 69.90776 Acc -> 0.9517245\n",
      "epoch 172 E loss -> 69.52612 Acc -> 0.95187235\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 173 E loss -> 69.52475 Acc -> 0.95186454\n",
      "epoch 174 E loss -> 69.23489 Acc -> 0.9520267\n",
      "epoch 175 E loss -> 69.364845 Acc -> 0.952027\n",
      "epoch 176 E loss -> 69.22577 Acc -> 0.95206624\n",
      "epoch 177 E loss -> 69.44859 Acc -> 0.95191854\n",
      "epoch 178 E loss -> 69.58396 Acc -> 0.9518734\n",
      "epoch 179 E loss -> 69.56943 Acc -> 0.9518701\n",
      "epoch 180 E loss -> 69.51563 Acc -> 0.9519057\n",
      "epoch 181 E loss -> 69.7011 Acc -> 0.95176667\n",
      "epoch 182 E loss -> 69.57831 Acc -> 0.95191574\n",
      "epoch 183 E loss -> 69.18797 Acc -> 0.9520786\n",
      "epoch 184 E loss -> 68.88336 Acc -> 0.95228595\n",
      "epoch 185 E loss -> 69.22283 Acc -> 0.95206594\n",
      "epoch 186 E loss -> 69.03972 Acc -> 0.95220673\n",
      "epoch 187 E loss -> 69.29314 Acc -> 0.95202\n",
      "epoch 188 E loss -> 69.0499 Acc -> 0.952196\n",
      "epoch 189 E loss -> 69.488976 Acc -> 0.9519211\n",
      "epoch 190 E loss -> 69.173134 Acc -> 0.95209694\n",
      "epoch 191 E loss -> 69.37706 Acc -> 0.95196676\n",
      "epoch 192 E loss -> 69.30583 Acc -> 0.95200276\n",
      "epoch 193 E loss -> 69.173515 Acc -> 0.95212775\n",
      "epoch 194 E loss -> 69.13559 Acc -> 0.9521422\n",
      "epoch 195 E loss -> 69.36716 Acc -> 0.95199496\n",
      "epoch 196 E loss -> 69.0015 Acc -> 0.95218724\n",
      "epoch 197 E loss -> 68.94534 Acc -> 0.95228547\n",
      "epoch 198 E loss -> 68.91859 Acc -> 0.9523099\n",
      "epoch 199 E loss -> 68.99288 Acc -> 0.9522523\n",
      "epoch 200 E loss -> 68.53363 Acc -> 0.9525124\n",
      "epoch 201 E loss -> 69.13185 Acc -> 0.9522131\n",
      "epoch 202 E loss -> 69.29632 Acc -> 0.9520324\n",
      "epoch 203 E loss -> 69.058975 Acc -> 0.9522251\n",
      "epoch 204 E loss -> 68.85743 Acc -> 0.9523164\n",
      "epoch 205 E loss -> 69.12485 Acc -> 0.95211667\n",
      "epoch 206 E loss -> 69.03627 Acc -> 0.95223856\n",
      "epoch 207 E loss -> 69.114845 Acc -> 0.9521625\n",
      "epoch 208 E loss -> 68.6859 Acc -> 0.95245254\n",
      "epoch 209 E loss -> 68.69379 Acc -> 0.95240885\n",
      "epoch 210 E loss -> 68.585556 Acc -> 0.9525122\n",
      "epoch 211 E loss -> 68.66612 Acc -> 0.9523967\n",
      "epoch 212 E loss -> 68.88712 Acc -> 0.95229584\n",
      "epoch 213 E loss -> 68.690674 Acc -> 0.952473\n",
      "epoch 214 E loss -> 69.789566 Acc -> 0.9517083\n",
      "epoch 215 E loss -> 68.77095 Acc -> 0.95237416\n",
      "epoch 216 E loss -> 69.04278 Acc -> 0.95223397\n",
      "epoch 217 E loss -> 68.60753 Acc -> 0.95244706\n",
      "epoch 218 E loss -> 68.8339 Acc -> 0.9523568\n",
      "epoch 219 E loss -> 68.40647 Acc -> 0.9526399\n",
      "epoch 220 E loss -> 68.65548 Acc -> 0.9524356\n",
      "epoch 221 E loss -> 68.70138 Acc -> 0.9524547\n",
      "epoch 222 E loss -> 68.65892 Acc -> 0.9524338\n",
      "epoch 223 E loss -> 69.14625 Acc -> 0.95212626\n",
      "epoch 224 E loss -> 68.71643 Acc -> 0.9524043\n",
      "epoch 225 E loss -> 68.46394 Acc -> 0.95258933\n",
      "epoch 226 E loss -> 68.72997 Acc -> 0.95243454\n",
      "epoch 227 E loss -> 68.746925 Acc -> 0.95236415\n",
      "epoch 228 E loss -> 68.614685 Acc -> 0.95245045\n",
      "epoch 229 E loss -> 68.79893 Acc -> 0.9523627\n",
      "epoch 230 E loss -> 68.9363 Acc -> 0.95222664\n",
      "epoch 231 E loss -> 68.72823 Acc -> 0.95241565\n",
      "epoch 232 E loss -> 68.5503 Acc -> 0.9525409\n",
      "epoch 233 E loss -> 68.84508 Acc -> 0.95228875\n",
      "epoch 234 E loss -> 69.05017 Acc -> 0.95220613\n",
      "epoch 235 E loss -> 68.47139 Acc -> 0.95260566\n",
      "epoch 236 E loss -> 68.57451 Acc -> 0.9525072\n",
      "epoch 237 E loss -> 68.79827 Acc -> 0.9523339\n",
      "epoch 238 E loss -> 68.32679 Acc -> 0.95266604\n",
      "epoch 239 E loss -> 68.4225 Acc -> 0.95262045\n",
      "epoch 240 E loss -> 68.8409 Acc -> 0.9523841\n",
      "epoch 241 E loss -> 68.57216 Acc -> 0.95246136\n",
      "epoch 242 E loss -> 68.342766 Acc -> 0.9526631\n",
      "epoch 243 E loss -> 68.82989 Acc -> 0.95233774\n",
      "epoch 244 E loss -> 68.63529 Acc -> 0.9524482\n",
      "epoch 245 E loss -> 68.36977 Acc -> 0.9526438\n",
      "epoch 246 E loss -> 68.55526 Acc -> 0.9525283\n",
      "epoch 247 E loss -> 68.48012 Acc -> 0.9525672\n",
      "epoch 248 E loss -> 68.396416 Acc -> 0.95254314\n",
      "epoch 249 E loss -> 68.52422 Acc -> 0.9525216\n",
      "epoch 250 E loss -> 68.306114 Acc -> 0.95266676\n",
      "Training stoped, last epoch reached ... \n",
      "Time:  14913.264678824693\n"
     ]
    }
   ],
   "source": [
    "num_epochs=25\n",
    "model = WGAN2D_MPS(summary = True, input_shape = (n_rows, n_cols, n_chns),SpectralNormalization=True)\n",
    "start = timeit.default_timer()\n",
    "model.train(x_train , epochs=num_epochs)\n",
    "stop = timeit.default_timer()\n",
    "print('Time: ', stop - start) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib  import pylab as plt\n",
    "plt.subplot(1,2,1)\n",
    "plt.plot(model.history['loss_g'])\n",
    "plt.subplot(1,2,2)\n",
    "plt.plot(model.history['loss_d'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "PlotDataAE(np.argmax(x_train[:20], axis=-1), model.Decoder(x_train[:20], True), digit_size=(n_rows, n_cols), Only_Result=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x144 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_random = np.random.normal(0, 1, size=(20, 500))\n",
    "model_plot(x_random, model, num_plots = 6, mode = \"gen\", figsize = (60,60), type_model = \"CycleGAN\" )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "'return' outside function (<ipython-input-12-9b32c1431b19>, line 1)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-12-9b32c1431b19>\"\u001b[0;36m, line \u001b[0;32m1\u001b[0m\n\u001b[0;31m    return\u001b[0m\n\u001b[0m          ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m 'return' outside function\n"
     ]
    }
   ],
   "source": [
    "return"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. AlphaGAN_MPS Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_epochs = 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "start = timeit.default_timer()\n",
    "model = AlphaGAN_MPS(summary = True, input_shape = (n_rows, n_cols, n_chns),alpha=100,\n",
    "                    SpectralNormalization=True,saving_path='/share/GeoFacies/GeoFacies_DL/')\n",
    "model.train(x_train , epochs=num_epochs, num_plots = 10)\n",
    "stop = timeit.default_timer()\n",
    "print('Time: ', stop - start) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "PlotDataAE(np.argmax(x_train[:20], axis=-1), model.Decoder(x_train[:20], True), digit_size=(n_rows, n_cols), Only_Result=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_random = np.random.normal(0, 1, size=(20, 500))\n",
    "model_plot(x_random, model, num_plots = 4, mode = \"gen\", figsize = (60,60), type_model = \"CycleGAN\" )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4.  CycleGAN Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start = timeit.default_timer()\n",
    "model = CycleGAN_MPS(summary = True, input_shape = (n_rows, n_cols, n_chns),batch_size=2,\n",
    "                     epsilon=0.5,SpectralNormalization=True,saving_path='/share/GeoFacies/GeoFacies_DL/')\n",
    "model.train(x_train , epochs=num_epochs)\n",
    "stop = timeit.default_timer()\n",
    "print('Time: ', stop - start) "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}