Predict_class_file_written_from_scratch_Numpy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# In this file I am trying to test trained model/ classify image by writing a code only using Numpy. No Keras API has used here. Main part will start by extracting the weight from model file. For loading model.h5 file Keras has used. Preprocessing also has done using some other scikit learn."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# This file will work for multi convolution filter."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocessing the image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import sys\n",
    "from skimage import io, color, exposure, transform\n",
    "from skimage.color import rgb2gray\n",
    "#from sklearn.cross_validation import train_test_split\n",
    "from sklearn.model_selection import train_test_split  #it came from update scikit learn. https://stackoverflow.com/questions/40704484/importerror-no-module-named-model-selection\n",
    "import os\n",
    "import glob\n",
    "import h5py\n",
    "\n",
    "from matplotlib import pyplot as plt\n",
    "#%matplotlib inline\n",
    "\n",
    "# NUM_CLASSES = 9 #Used class for the training\n",
    "# IMG_SIZE = 48 #required size. This size has also maintained during training. User defined value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import sys\n",
    "#from skimage import io, color, exposure, transform\n",
    "#from skimage.color import rgb2gray\n",
    "#from sklearn.cross_validation import train_test_split\n",
    "#from sklearn.model_selection import train_test_split  #it came from update scikit learn. https://stackoverflow.com/questions/40704484/importerror-no-module-named-model-selection\n",
    "import os, json\n",
    "import glob\n",
    "import h5py\n",
    "import time\n",
    "import cv2\n",
    "\n",
    "with open('variable_config.json', 'r') as f:\n",
    "    config = json.load(f)\n",
    "\n",
    "\n",
    "NUM_CLASSES = config['DEFAULT']['num_class']\n",
    "IMG_SIZE = config['DEFAULT']['img_size']\n",
    "number_filter = config['DEFAULT']['number_filter']\n",
    "filter_size = config['DEFAULT']['filter_size']\n",
    "img_depth = config['DEFAULT']['img_depth']\n",
    "img_type = config['DEFAULT']['img_type']\n",
    "epochs = config['DEFAULT']['epochs']\n",
    "batch_size = config['DEFAULT']['batch_size']\n",
    "train_image_path = config['DEFAULT']['train_image_path']\n",
    "test_image_path = config['DEFAULT']['test_image_path']\n",
    "learning_model_path = config['DEFAULT']['learning_model_path']\n",
    "validation_split = config['DEFAULT']['validation_split']\n",
    "\n",
    "#from matplotlib import pyplot as plt\n",
    "#%matplotlib inline\n",
    "\n",
    "# NUM_CLASSES = 9 #Used class for the training\n",
    "# IMG_SIZE = 8 #required size. This size has also maintained during training. User defined value\n",
    "# number_filter=1\n",
    "total_time=0.000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_regularizer': None, 'kernel_size': (3, 3), 'name': 'conv2d_1', 'dtype': 'float32', 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'bias_regularizer': None, 'batch_input_shape': (None, 1, 48, 48), 'padding': 'same', 'use_bias': True, 'kernel_constraint': None, 'filters': 32} \n",
      "\n",
      "h==  [array([[[[-3.09373170e-01,  3.35247469e+00, -8.17258190e-03,\n",
      "          -1.52699903e-01, -3.92200202e-02, -4.09431569e-02,\n",
      "          -6.50194585e-02,  3.95465112e+00, -2.88339257e-02,\n",
      "           9.65168417e-01,  2.26394725e+00,  9.19752643e-02,\n",
      "           1.04308200e+00,  5.02101851e+00, -1.16733693e-01,\n",
      "          -6.56417534e-02,  2.01192570e+00,  9.72868279e-02,\n",
      "          -1.16467252e-01,  1.52933329e-01,  1.46493137e-01,\n",
      "           2.17062030e-02,  5.51404692e-02,  3.53696299e+00,\n",
      "           4.58308786e-01, -1.42728373e-01, -4.09316383e-02,\n",
      "           3.49415421e+00, -9.86798704e-02, -7.59547353e-02,\n",
      "           3.87289023e+00,  2.35459709e+00]],\n",
      "\n",
      "        [[-3.38641018e-01,  3.43956995e+00, -4.37568873e-02,\n",
      "           2.05426086e-02, -6.38462678e-02, -8.28460529e-02,\n",
      "           2.52780113e-02,  4.17110682e+00, -1.41278327e-01,\n",
      "           8.75337839e-01,  2.24116945e+00, -5.90169020e-02,\n",
      "           1.12359333e+00,  4.99869680e+00, -1.12865284e-01,\n",
      "          -1.11755384e-02,  1.86557484e+00, -2.17131153e-02,\n",
      "           1.07350290e-01, -2.62478907e-02, -1.33075669e-01,\n",
      "           1.76684577e-02, -1.00192830e-01,  3.45771503e+00,\n",
      "           5.16355932e-01, -2.03536615e-01, -1.34889647e-01,\n",
      "           3.60002398e+00, -1.41094133e-01, -1.30550429e-01,\n",
      "           3.80404592e+00,  2.31310749e+00]],\n",
      "\n",
      "        [[-2.69378066e-01,  3.29916310e+00,  1.00698866e-01,\n",
      "          -1.02891922e-02,  2.16074809e-02, -9.02320668e-02,\n",
      "          -1.16692565e-01,  4.02014494e+00,  3.62323821e-02,\n",
      "           9.14092720e-01,  2.20846891e+00, -8.22363496e-02,\n",
      "           9.95436549e-01,  4.87846518e+00, -1.24915928e-01,\n",
      "           9.53978822e-02,  1.73148465e+00, -9.55658965e-03,\n",
      "           1.82193909e-02,  4.85175289e-02,  3.25670429e-02,\n",
      "           1.53056430e-02,  9.67345014e-02,  3.40354753e+00,\n",
      "           4.47713226e-01, -1.65807560e-01, -2.48249155e-02,\n",
      "           3.39432311e+00, -1.05185686e-02, -3.40212919e-02,\n",
      "           3.68116856e+00,  2.26339674e+00]]],\n",
      "\n",
      "\n",
      "       [[[-1.42402902e-01,  3.27248263e+00,  4.66080382e-02,\n",
      "          -5.47038130e-02, -1.53161272e-01, -5.82188144e-02,\n",
      "           7.87916705e-02,  3.95243692e+00, -8.48340243e-03,\n",
      "           1.05669081e+00,  2.10688400e+00, -1.34589359e-01,\n",
      "           8.76845360e-01,  4.82489967e+00,  5.96768111e-02,\n",
      "           4.66563962e-02,  1.94212937e+00,  1.14101104e-01,\n",
      "           4.62026633e-02,  3.23026255e-02,  1.27386332e-01,\n",
      "          -1.30084738e-01,  3.49206328e-02,  3.36314869e+00,\n",
      "           5.77509284e-01, -2.66903043e-02, -2.27216091e-02,\n",
      "           3.47769499e+00, -4.15516235e-02,  7.84290880e-02,\n",
      "           3.85823417e+00,  2.40642476e+00]],\n",
      "\n",
      "        [[-1.47521928e-01,  3.48220563e+00, -1.10951774e-02,\n",
      "          -3.40171345e-02, -8.27020258e-02,  2.29802541e-02,\n",
      "          -1.33519992e-01,  4.10947704e+00, -7.26879463e-02,\n",
      "           1.03724945e+00,  2.14315200e+00,  8.78517702e-02,\n",
      "           1.07152927e+00,  5.00246525e+00, -8.25458243e-02,\n",
      "          -1.28494158e-01,  1.76456392e+00, -1.22840561e-01,\n",
      "          -9.87263620e-02, -5.86965261e-03, -1.00361601e-01,\n",
      "          -3.60378958e-02, -3.16587724e-02,  3.29109192e+00,\n",
      "           3.97692233e-01, -1.55687690e-01,  7.56319240e-02,\n",
      "           3.53860712e+00,  2.89036869e-03,  7.64918998e-02,\n",
      "           3.74333215e+00,  2.19153237e+00]],\n",
      "\n",
      "        [[-1.13230884e-01,  3.28517079e+00,  1.48765622e-02,\n",
      "           6.80713132e-02, -8.74133501e-03,  3.28972042e-02,\n",
      "           1.30822301e-01,  4.08517122e+00, -2.98831090e-02,\n",
      "           9.31659162e-01,  2.30330276e+00, -1.13465577e-01,\n",
      "           1.04079044e+00,  4.97926092e+00,  4.46644910e-02,\n",
      "          -6.95486292e-02,  1.81070042e+00,  2.16994155e-02,\n",
      "          -1.30475745e-01, -6.84672445e-02,  2.04589795e-02,\n",
      "          -1.00446366e-01, -4.77962270e-02,  3.21759295e+00,\n",
      "           3.13956141e-01,  3.22208442e-02, -1.23100474e-01,\n",
      "           3.35459828e+00, -2.12714523e-02,  8.00873414e-02,\n",
      "           3.55919170e+00,  2.13463354e+00]]],\n",
      "\n",
      "\n",
      "       [[[-2.26866543e-01,  3.39521170e+00, -6.72599450e-02,\n",
      "           4.48007286e-02, -1.82539582e-01,  1.98313389e-02,\n",
      "           1.07447848e-01,  3.99482274e+00,  4.31856215e-02,\n",
      "           8.05238843e-01,  2.12024355e+00, -3.20941322e-02,\n",
      "           9.62297499e-01,  4.94219542e+00, -9.95145589e-02,\n",
      "          -1.39926538e-01,  1.90619969e+00, -9.37294587e-02,\n",
      "           5.46604730e-02, -2.17119232e-02, -1.29598424e-01,\n",
      "           9.41729695e-02, -9.99556780e-02,  3.43063521e+00,\n",
      "           2.99992830e-01, -1.25056831e-03, -1.38971200e-02,\n",
      "           3.28418851e+00,  1.11651063e-01, -1.16201900e-01,\n",
      "           3.64895511e+00,  2.14763641e+00]],\n",
      "\n",
      "        [[-3.19729000e-01,  3.13530016e+00, -1.23586863e-01,\n",
      "          -5.17112538e-02,  4.74050529e-02, -5.44978157e-02,\n",
      "          -9.57548171e-02,  3.91495299e+00, -1.18388310e-01,\n",
      "           8.12726080e-01,  2.09675884e+00, -6.33397698e-02,\n",
      "           9.58534420e-01,  4.83660984e+00, -1.39002670e-02,\n",
      "           8.78791735e-02,  1.95243657e+00, -2.67206207e-02,\n",
      "          -4.49119024e-02, -3.40639129e-02, -1.30016267e-01,\n",
      "          -8.57137293e-02, -8.34661499e-02,  3.37403655e+00,\n",
      "           5.54753780e-01, -5.16405441e-02, -9.30356421e-03,\n",
      "           3.39549541e+00,  1.07618041e-01, -1.24180317e-01,\n",
      "           3.73351955e+00,  2.28144336e+00]],\n",
      "\n",
      "        [[-1.55600101e-01,  3.32652211e+00,  6.54247105e-02,\n",
      "          -1.51523337e-01,  8.50881413e-02,  3.04461997e-02,\n",
      "          -8.85324851e-02,  3.89690208e+00, -3.58887692e-03,\n",
      "           8.93082559e-01,  2.20282936e+00, -2.61048358e-02,\n",
      "           9.17856395e-01,  4.70491791e+00, -7.18671605e-02,\n",
      "           5.12253679e-02,  1.91491842e+00, -1.08022571e-01,\n",
      "          -1.18728414e-01, -7.95996934e-02, -1.39560878e-01,\n",
      "          -1.25697985e-01, -7.29252994e-02,  3.31700110e+00,\n",
      "           4.51437563e-01, -1.54678985e-01,  1.51565550e-02,\n",
      "           3.43135619e+00, -1.08369745e-01, -3.11297365e-02,\n",
      "           3.55899215e+00,  2.09235477e+00]]]], dtype=float32), array([-7.1496415e-01,  7.3358831e+00, -1.4526361e-03, -7.4612796e-02,\n",
      "       -9.2530310e-02, -6.3542255e-05, -5.5468595e-04,  8.8747873e+00,\n",
      "       -9.3253405e-08,  2.0151806e+00,  4.8058496e+00,  3.6226791e-06,\n",
      "        2.2141354e+00,  1.0806567e+01,  1.3514644e-07,  8.2985498e-03,\n",
      "        4.1705885e+00, -5.2090161e-03, -5.8526076e-02,  5.7045165e-02,\n",
      "        1.0794052e-02, -4.0294285e-06,  2.5901417e-03,  7.4188881e+00,\n",
      "        1.0124300e+00, -2.2153082e-01,  1.0213382e-06,  7.6324244e+00,\n",
      "        3.1386816e-03, -1.3744909e-05,  8.1925831e+00,  4.9539576e+00],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_size': (3, 3), 'name': 'conv2d_2', 'bias_regularizer': None, 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'filters': 64, 'use_bias': True, 'padding': 'valid', 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[[[-0.05204006, -0.05165289, -0.02130643, ...,  0.04129837,\n",
      "          -0.00879113, -0.02419413],\n",
      "         [ 0.01962756, -0.08165328, -0.01954962, ...,  0.07213595,\n",
      "           0.04712817, -0.03786406],\n",
      "         [-0.02850312,  0.02806563, -0.06285219, ...,  0.07376888,\n",
      "          -0.04521652,  0.01391444],\n",
      "         ...,\n",
      "         [-0.07525748, -0.06515543,  0.03387736, ..., -0.02229894,\n",
      "           0.01500255,  0.032488  ],\n",
      "         [ 0.07506266, -0.0468437 , -0.05870821, ...,  0.22484425,\n",
      "           0.0101591 ,  0.01144099],\n",
      "         [ 0.04158848,  0.02468054, -0.08089433, ...,  0.14432907,\n",
      "          -0.01863596,  0.07467945]],\n",
      "\n",
      "        [[-0.02777869, -0.06385736,  0.00736684, ..., -0.05597709,\n",
      "          -0.00991438, -0.05432329],\n",
      "         [ 0.06487349,  0.00703766,  0.02141744, ...,  0.09425829,\n",
      "          -0.01531525,  0.01668831],\n",
      "         [-0.02936508, -0.03000946, -0.06557753, ...,  0.08082409,\n",
      "           0.08236921,  0.04473926],\n",
      "         ...,\n",
      "         [ 0.07644379,  0.02159083, -0.04895004, ..., -0.08053441,\n",
      "          -0.03068189, -0.01663853],\n",
      "         [ 0.05993341, -0.02070566,  0.04763356, ...,  0.19023229,\n",
      "           0.07225645,  0.03356617],\n",
      "         [-0.0368784 ,  0.06257592,  0.01901006, ...,  0.17020807,\n",
      "          -0.00541168, -0.0290654 ]],\n",
      "\n",
      "        [[-0.02885871, -0.0822996 ,  0.03781106, ..., -0.07329991,\n",
      "           0.03886244,  0.0424326 ],\n",
      "         [ 0.02977155, -0.02744321, -0.00708452, ...,  0.15108387,\n",
      "          -0.01772066,  0.01005678],\n",
      "         [-0.00044154, -0.04653879,  0.03884913, ..., -0.0365649 ,\n",
      "          -0.04969597,  0.0114736 ],\n",
      "         ...,\n",
      "         [ 0.05407207, -0.0018126 , -0.07608896, ..., -0.07018664,\n",
      "           0.04888343,  0.01678826],\n",
      "         [ 0.08510803,  0.03470738, -0.0164835 , ...,  0.07619903,\n",
      "           0.01233336, -0.04497018],\n",
      "         [ 0.04287382, -0.05393887, -0.04150967, ...,  0.05420488,\n",
      "           0.06963326, -0.01986465]]],\n",
      "\n",
      "\n",
      "       [[[ 0.04134207,  0.03815904,  0.0381147 , ..., -0.07572479,\n",
      "          -0.06318134,  0.00510881],\n",
      "         [-0.07124875, -0.01556673, -0.04120636, ...,  0.11915717,\n",
      "          -0.07558798, -0.04000577],\n",
      "         [ 0.02805041,  0.07942034, -0.04198608, ...,  0.0319232 ,\n",
      "          -0.07240955,  0.03590528],\n",
      "         ...,\n",
      "         [-0.00138706, -0.02511344,  0.06564087, ...,  0.07864954,\n",
      "           0.01936582, -0.0283858 ],\n",
      "         [-0.02185424,  0.07687662,  0.04550012, ...,  0.17569803,\n",
      "          -0.01451143, -0.04049436],\n",
      "         [ 0.06819932,  0.02264786, -0.07586806, ...,  0.15840745,\n",
      "          -0.0183832 , -0.07429119]],\n",
      "\n",
      "        [[-0.06440253, -0.00336273, -0.01056222, ...,  0.06168346,\n",
      "          -0.02129117, -0.04787959],\n",
      "         [-0.07708144, -0.01004121, -0.01063206, ...,  0.21043259,\n",
      "          -0.06090438,  0.02357454],\n",
      "         [ 0.00373329,  0.02401765,  0.04861209, ...,  0.0638748 ,\n",
      "           0.04358368, -0.05829545],\n",
      "         ...,\n",
      "         [-0.00140493,  0.0655726 ,  0.05234901, ..., -0.00204383,\n",
      "          -0.02801666, -0.01158816],\n",
      "         [ 0.03968832,  0.02639259, -0.02034978, ...,  0.10193684,\n",
      "          -0.01240591,  0.0579825 ],\n",
      "         [ 0.022399  , -0.04219379,  0.03615176, ...,  0.14627147,\n",
      "          -0.04789411,  0.06518719]],\n",
      "\n",
      "        [[-0.05914979, -0.07321398,  0.06312644, ...,  0.02958548,\n",
      "          -0.06995886,  0.02357111],\n",
      "         [ 0.0253478 , -0.02669057,  0.01306575, ...,  0.14737561,\n",
      "          -0.0169927 ,  0.07219355],\n",
      "         [-0.0757373 , -0.06483756, -0.04654388, ..., -0.06052949,\n",
      "           0.00384367, -0.04989797],\n",
      "         ...,\n",
      "         [-0.00054469,  0.00647807, -0.06886851, ..., -0.0132888 ,\n",
      "           0.06065769,  0.03554606],\n",
      "         [-0.06951624, -0.04210459, -0.02834794, ...,  0.16221479,\n",
      "           0.03947905, -0.04280178],\n",
      "         [-0.04637346, -0.03347175,  0.02804274, ...,  0.14752898,\n",
      "           0.07627639,  0.0828414 ]]],\n",
      "\n",
      "\n",
      "       [[[ 0.06091116,  0.0290173 , -0.06772208, ...,  0.05274678,\n",
      "           0.02999458, -0.02569036],\n",
      "         [-0.00796046, -0.01440886, -0.10290957, ...,  0.08850541,\n",
      "           0.01257788, -0.02862517],\n",
      "         [ 0.05708287, -0.05838758, -0.03745097, ...,  0.07215326,\n",
      "          -0.00915252, -0.07768777],\n",
      "         ...,\n",
      "         [-0.08260223, -0.07004255, -0.00827644, ...,  0.05992608,\n",
      "          -0.00426917,  0.03079524],\n",
      "         [-0.00979742, -0.00702913,  0.01432021, ...,  0.21735717,\n",
      "          -0.06583554, -0.04594952],\n",
      "         [-0.01055741,  0.07214206, -0.02928031, ...,  0.04569571,\n",
      "           0.04932301,  0.01100751]],\n",
      "\n",
      "        [[-0.07976939, -0.02347391, -0.01908376, ...,  0.07965762,\n",
      "          -0.03716754, -0.05397699],\n",
      "         [ 0.07600733,  0.04054283, -0.04872248, ...,  0.07846758,\n",
      "          -0.02946095,  0.06895053],\n",
      "         [-0.06931357, -0.06794364,  0.04540797, ..., -0.02285864,\n",
      "          -0.05981857, -0.06923971],\n",
      "         ...,\n",
      "         [-0.05823054, -0.05389907, -0.07242208, ..., -0.01284694,\n",
      "          -0.02703059, -0.02408369],\n",
      "         [ 0.01476161,  0.04103174,  0.03577217, ...,  0.12286241,\n",
      "          -0.0174335 , -0.01031537],\n",
      "         [-0.01661594, -0.06755605, -0.01095899, ...,  0.09642462,\n",
      "          -0.01623811, -0.02992911]],\n",
      "\n",
      "        [[-0.05559262,  0.02876953, -0.04212078, ..., -0.04249275,\n",
      "           0.00373131,  0.07624796],\n",
      "         [-0.00191511, -0.05411363, -0.04351168, ...,  0.13210845,\n",
      "          -0.03375446,  0.06450593],\n",
      "         [ 0.0524465 , -0.04597639, -0.06041453, ...,  0.01997201,\n",
      "           0.08145661, -0.00470754],\n",
      "         ...,\n",
      "         [ 0.06361047, -0.02657112,  0.01361472, ..., -0.03875402,\n",
      "          -0.0152473 , -0.00970734],\n",
      "         [ 0.05750667,  0.02277932,  0.00634411, ...,  0.08122636,\n",
      "          -0.06142933, -0.08086754],\n",
      "         [ 0.00886445, -0.0033918 , -0.04336973, ...,  0.11781656,\n",
      "           0.02228854,  0.05566244]]]], dtype=float32), array([ 1.0676484e-02,  1.3947596e-04, -6.4672157e-02, -1.7107128e-01,\n",
      "        1.1251543e+00,  1.8198166e-02, -3.0279843e-05, -7.9561054e-04,\n",
      "       -2.8376007e-03,  4.6418095e+00, -1.2060591e-02,  2.5354609e+00,\n",
      "       -2.2681928e-03,  1.1892143e+00,  1.8810383e+00, -6.9878340e-02,\n",
      "       -9.3046106e-02,  1.8429861e-03,  1.8657899e+00, -6.6675995e-05,\n",
      "       -7.8855531e-04,  3.8413140e-01,  4.0119082e-01, -1.6635694e-03,\n",
      "       -8.1626035e-02,  1.4245827e+00,  1.3645256e+00,  4.8152205e-01,\n",
      "       -7.4503431e-04,  1.6943310e+00, -6.5555826e-02,  2.5859704e+00,\n",
      "       -7.1693562e-02,  8.5787237e-01,  1.1259266e+00,  1.9721105e+00,\n",
      "        1.7575572e-04,  2.5004765e-02,  2.1256767e-01,  1.5963495e-03,\n",
      "       -1.6044352e-02,  9.1682541e-01,  2.1558747e+00,  1.0506641e+00,\n",
      "        1.0472208e+00,  9.2046134e-02,  2.5766892e+00,  1.6195865e+00,\n",
      "        7.5946212e-01, -8.3845033e-04,  3.8207704e-03,  1.1836456e-03,\n",
      "       -2.0777984e-04, -3.1356536e-02,  1.8093303e-02,  1.1295602e+00,\n",
      "        2.4952052e+00, -1.0168262e-02,  6.5788319e-03,  2.9316711e-01,\n",
      "        1.8604883e+00,  3.9992034e-01,  3.3822312e-04,  6.1309133e-03],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'padding': 'valid', 'data_format': 'channels_first', 'strides': (2, 2), 'pool_size': (2, 2), 'name': 'max_pooling2d_1', 'trainable': True} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'rate': 0.2, 'name': 'dropout_1', 'seed': None, 'trainable': True, 'noise_shape': None} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_regularizer': None, 'kernel_size': (3, 3), 'name': 'conv2d_3', 'dtype': 'float32', 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'bias_regularizer': None, 'batch_input_shape': (None, 3, 48, 48), 'padding': 'same', 'use_bias': True, 'kernel_constraint': None, 'filters': 128} \n",
      "\n",
      "h==  [array([[[[-8.56497604e-03,  5.61562658e-04, -3.67604047e-02, ...,\n",
      "           1.55121563e-02, -3.12413611e-02, -4.27748971e-02],\n",
      "         [ 5.40847331e-02, -3.64951454e-02, -1.18955243e-02, ...,\n",
      "           4.80896048e-02, -5.58170397e-03,  5.05787767e-02],\n",
      "         [-3.78410220e-02, -3.29351611e-02,  4.60368060e-02, ...,\n",
      "          -2.34350823e-02, -9.36319120e-03, -1.80610269e-03],\n",
      "         ...,\n",
      "         [ 2.81477789e-03, -7.16403732e-03, -5.42641170e-02, ...,\n",
      "           2.06866469e-02,  5.19880978e-03, -5.73179722e-02],\n",
      "         [ 2.96088755e-02,  5.17552271e-02, -1.08596291e-02, ...,\n",
      "          -1.72570962e-02, -5.64877614e-02,  9.23595671e-03],\n",
      "         [ 1.42810140e-02,  5.59962690e-02,  7.85762910e-03, ...,\n",
      "           5.55486530e-02, -2.74579823e-02, -3.26812901e-02]],\n",
      "\n",
      "        [[-3.31830531e-02, -1.96958520e-03, -2.40556058e-02, ...,\n",
      "           4.40494381e-02, -4.95057330e-02, -2.55751293e-02],\n",
      "         [ 6.30501891e-03,  2.33699079e-03, -2.51151901e-02, ...,\n",
      "           4.53078412e-02,  8.67847726e-03, -4.45467047e-02],\n",
      "         [ 9.52317473e-03,  1.04531401e-03,  7.32779503e-04, ...,\n",
      "          -4.27087136e-02,  1.16519472e-02, -5.00686280e-02],\n",
      "         ...,\n",
      "         [-4.77951299e-03,  4.54092547e-02,  4.87257093e-02, ...,\n",
      "           2.87664644e-02, -3.88486013e-02, -7.60680530e-03],\n",
      "         [-4.66634259e-02,  5.58948703e-03,  3.00396569e-02, ...,\n",
      "           1.15853595e-02,  5.87889776e-02, -3.73586640e-02],\n",
      "         [-5.04725911e-02,  4.39940430e-02,  1.22656692e-02, ...,\n",
      "          -3.20886113e-02, -4.37665842e-02,  8.52458458e-03]],\n",
      "\n",
      "        [[ 5.28070778e-02,  5.29345721e-02, -5.64656928e-02, ...,\n",
      "           6.61602058e-03,  5.33085614e-02, -1.25962291e-02],\n",
      "         [-2.03019157e-02,  4.64225411e-02,  4.30170707e-02, ...,\n",
      "           1.14197526e-02,  3.83551754e-02,  1.41884200e-02],\n",
      "         [ 4.17090692e-02,  2.45658681e-02,  3.53239514e-02, ...,\n",
      "           1.56397708e-02, -5.30567467e-02, -1.32894590e-02],\n",
      "         ...,\n",
      "         [-4.81143184e-02,  3.00060418e-02,  4.12782729e-02, ...,\n",
      "           7.15653673e-02,  2.51112022e-02,  3.69498739e-04],\n",
      "         [-1.17935427e-02,  3.82208340e-02,  1.77526586e-02, ...,\n",
      "          -4.45599854e-02,  5.62933348e-02,  4.94506061e-02],\n",
      "         [ 6.58062473e-03,  2.31484622e-02, -4.41938303e-02, ...,\n",
      "          -1.98267922e-02, -7.98387732e-03,  2.50692423e-02]]],\n",
      "\n",
      "\n",
      "       [[[ 4.09287289e-02,  4.42005247e-02,  3.26497294e-02, ...,\n",
      "           3.82557362e-02,  3.08349412e-02,  1.28356200e-02],\n",
      "         [ 4.27431166e-02, -4.29794937e-02, -2.78967023e-02, ...,\n",
      "           4.54305075e-02,  1.27607305e-02, -5.33137247e-02],\n",
      "         [-3.93980630e-02,  6.95443749e-02, -2.98472606e-02, ...,\n",
      "           2.19834056e-02, -3.54705714e-02, -4.26973477e-02],\n",
      "         ...,\n",
      "         [-5.14154881e-02, -5.73669747e-03,  3.68105993e-02, ...,\n",
      "           7.31929764e-02, -5.06636826e-03, -4.76002635e-04],\n",
      "         [ 1.09691992e-02,  2.77464483e-02,  4.74154130e-02, ...,\n",
      "           9.59882513e-03, -4.19544540e-02, -4.87796627e-02],\n",
      "         [ 5.58819901e-03, -4.09580022e-03, -9.85565782e-03, ...,\n",
      "           5.46807162e-02, -3.48168872e-02,  2.18579341e-02]],\n",
      "\n",
      "        [[-3.03517487e-02, -1.15054995e-02, -5.49052320e-02, ...,\n",
      "          -1.06657753e-02,  5.22892512e-02,  2.78949570e-02],\n",
      "         [-4.36389223e-02,  3.17552052e-02,  4.40183543e-02, ...,\n",
      "           4.69691008e-02,  2.83811074e-02, -2.82623563e-02],\n",
      "         [-4.32212837e-03,  7.14966655e-02, -1.58531312e-02, ...,\n",
      "           3.25502641e-02,  3.12654525e-02, -1.91980600e-02],\n",
      "         ...,\n",
      "         [ 2.09486391e-03,  9.78006702e-03,  2.58281343e-02, ...,\n",
      "          -2.54942961e-02, -1.83781199e-02, -2.37881411e-02],\n",
      "         [-3.96632962e-02,  1.73119307e-02,  4.70068902e-02, ...,\n",
      "           5.39945029e-02,  1.97597239e-02, -4.54946458e-02],\n",
      "         [-9.75066610e-03,  4.41590808e-02, -7.60531006e-03, ...,\n",
      "           2.18122732e-02, -5.44445179e-02, -1.54614933e-02]],\n",
      "\n",
      "        [[ 2.01476850e-02,  1.79817546e-02,  4.79117315e-03, ...,\n",
      "          -1.81997884e-02,  3.44385803e-02, -9.96923354e-03],\n",
      "         [-2.66513345e-03, -5.80699332e-02, -7.83291180e-03, ...,\n",
      "           3.30720060e-02,  5.44335730e-02,  5.43489270e-02],\n",
      "         [ 3.92894819e-03,  6.53185770e-02, -3.42100896e-02, ...,\n",
      "           2.12700479e-02,  4.69992384e-02,  1.31079191e-02],\n",
      "         ...,\n",
      "         [ 2.13980395e-02,  5.14412522e-02, -1.09235216e-02, ...,\n",
      "           3.63412201e-02,  1.02014816e-03, -5.32819442e-02],\n",
      "         [ 2.81669665e-03,  3.83861661e-02,  2.20544003e-02, ...,\n",
      "          -3.96580249e-02,  5.72044961e-03,  3.25706340e-02],\n",
      "         [ 1.90849509e-02,  1.38751846e-02,  1.56804286e-02, ...,\n",
      "          -8.25620349e-03, -1.37062352e-02, -3.32760289e-02]]],\n",
      "\n",
      "\n",
      "       [[[ 5.42289764e-02, -3.73711698e-02, -1.98622532e-02, ...,\n",
      "           5.64656518e-02,  2.64977645e-02, -4.90737036e-02],\n",
      "         [ 4.75867391e-02,  3.74472886e-03,  3.69321294e-02, ...,\n",
      "           1.44723300e-02,  7.23686069e-03, -5.40938638e-02],\n",
      "         [ 3.80925611e-02,  8.92467517e-03, -5.64893410e-02, ...,\n",
      "          -4.46827151e-02,  1.36490567e-02,  2.31759828e-02],\n",
      "         ...,\n",
      "         [ 2.71339342e-02, -6.41744677e-03,  4.81470414e-02, ...,\n",
      "          -1.59821231e-02,  5.75894751e-02, -3.07335705e-02],\n",
      "         [-2.91845612e-02, -6.89445483e-03,  4.86618774e-05, ...,\n",
      "           6.23108307e-03,  5.83568327e-02,  2.83035561e-02],\n",
      "         [ 2.24248786e-02,  4.61707190e-02, -4.36994024e-02, ...,\n",
      "           5.34945242e-02, -2.74383537e-02,  5.40150777e-02]],\n",
      "\n",
      "        [[ 1.74030978e-02,  3.78884599e-02, -1.79386362e-02, ...,\n",
      "          -2.95447577e-02, -5.73204346e-02, -3.40909697e-02],\n",
      "         [-2.81326920e-02,  2.58116797e-02, -3.59594775e-03, ...,\n",
      "          -2.46675368e-02,  3.83071899e-02, -5.87977916e-02],\n",
      "         [ 2.66671311e-02,  5.90197556e-02,  5.81722595e-02, ...,\n",
      "           5.40210381e-02,  6.13482157e-03, -7.78836617e-03],\n",
      "         ...,\n",
      "         [ 5.57734482e-02,  1.17179919e-02,  5.30465953e-02, ...,\n",
      "           4.97980043e-02, -3.25556211e-02, -6.95723528e-03],\n",
      "         [ 2.97146533e-02,  1.02152764e-04,  6.92238007e-03, ...,\n",
      "           2.83672824e-03, -9.17391106e-03,  3.66704650e-02],\n",
      "         [-2.58351211e-04,  1.63809787e-02,  1.75313454e-03, ...,\n",
      "           3.71111706e-02,  6.65711286e-03,  7.01931713e-04]],\n",
      "\n",
      "        [[-2.75755506e-02,  3.49157192e-02,  3.83484177e-02, ...,\n",
      "          -4.12929170e-02,  1.36803370e-02,  3.14263217e-02],\n",
      "         [-2.46864930e-02,  2.47309785e-02,  3.48173268e-02, ...,\n",
      "          -2.76419669e-02, -6.25260873e-03,  5.28348261e-04],\n",
      "         [-9.13771149e-03, -4.29232977e-02,  1.02699613e-02, ...,\n",
      "           1.23108951e-02,  2.76674591e-02,  2.46068127e-02],\n",
      "         ...,\n",
      "         [ 2.27020821e-03,  5.39331604e-03,  2.10987814e-02, ...,\n",
      "           4.61563207e-02,  4.72704284e-02, -3.83685976e-02],\n",
      "         [ 3.82334813e-02,  3.74849997e-02,  4.27286252e-02, ...,\n",
      "           4.30062935e-02,  2.04790179e-02,  3.80995460e-02],\n",
      "         [ 1.11052291e-02, -3.01780440e-02,  7.09109241e-03, ...,\n",
      "          -1.16569148e-02, -4.06311303e-02, -7.46825361e-04]]]],\n",
      "      dtype=float32), array([ 1.3461341e-02,  3.3720168e-01, -1.3807687e-03,  3.7280080e-01,\n",
      "        1.0086772e+00,  1.7228922e-01,  2.5868070e-01,  1.6633296e-01,\n",
      "        2.9216364e-01,  7.5268926e-04,  1.3148509e-01,  8.3633251e-03,\n",
      "        6.7961566e-02, -3.1681517e-03,  3.9009026e-01,  2.0712766e-01,\n",
      "       -8.0546853e-04,  4.9782977e-01,  9.9069911e-01,  1.8191159e-01,\n",
      "       -2.3964990e-02,  1.2966971e+00,  6.6495836e-02,  7.8625977e-04,\n",
      "        1.4972254e-03,  2.5896934e-01,  4.8463252e-01,  4.7031641e-01,\n",
      "        3.3685748e-04,  2.6248759e-01,  1.2022375e-02,  1.7872636e-01,\n",
      "        4.3345362e-01,  3.0211427e-03,  7.0118625e-04,  3.8014871e-01,\n",
      "        5.7005662e-02,  4.7707227e-03,  1.1953996e-01,  4.9247167e-01,\n",
      "        8.3563089e-01, -2.4210261e-03,  1.0077387e-02,  4.0296462e-01,\n",
      "        9.0473217e-01,  6.0758807e-02,  5.8055639e-02,  6.0682293e-02,\n",
      "        4.6718670e-03,  2.9902208e-01,  1.3098240e-01,  5.4822274e-02,\n",
      "        1.6744401e-01,  5.8109947e-03,  5.5000955e-01,  4.3561966e-03,\n",
      "        8.0039787e-01, -2.1243775e-03,  1.7954050e-01,  9.2482293e-01,\n",
      "       -3.2501367e-03,  6.5721321e-01,  7.0936740e-03, -1.4757454e-02,\n",
      "        7.8564292e-01,  4.0293154e-01,  1.8397522e-01, -7.6289270e-03,\n",
      "        4.3353337e-01,  6.4238515e-03,  1.5585551e-01, -1.1768009e-02,\n",
      "        1.7485127e-01,  7.2229910e-03,  2.1174763e-04,  3.5899618e-01,\n",
      "        3.6773026e-02, -3.7057777e-03,  4.8248330e-03,  2.6996642e-01,\n",
      "        1.5223418e-04,  5.1123327e-01,  8.2457274e-02,  2.8541505e-01,\n",
      "        2.8457934e-01,  1.0627971e-02, -9.7343331e-04,  7.8777969e-02,\n",
      "        7.1350887e-02,  2.2391031e-02,  2.1471758e-01,  2.7853843e-02,\n",
      "        3.5808742e-01, -6.0412046e-03,  3.8503547e-04,  5.5289000e-01,\n",
      "        1.9384551e-01, -1.1262464e-03, -2.8418826e-03,  5.5260241e-01,\n",
      "        8.1732660e-04,  1.3180780e-03,  5.0734770e-01,  2.9355103e-02,\n",
      "        3.6403206e-01,  1.3797296e-02,  3.4212890e-01,  1.8671794e-01,\n",
      "        8.6102027e-01,  4.4554309e-03,  9.2304862e-01,  1.6956085e-01,\n",
      "       -5.3656260e-03,  3.5353804e-01,  1.8275807e-02,  9.8539060e-03,\n",
      "        6.0786720e-02,  5.7805026e-01,  1.9809283e-01,  6.4480671e-04,\n",
      "        1.8430935e-03,  7.6157689e-02,  6.7485715e-03,  6.7455244e-01,\n",
      "       -4.2220401e-03,  3.2192805e-01,  9.6803382e-02,  2.4483916e-03],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_size': (3, 3), 'name': 'conv2d_4', 'bias_regularizer': None, 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'filters': 256, 'use_bias': True, 'padding': 'valid', 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[[[-8.56832881e-03, -2.75387708e-03, -1.99029092e-02, ...,\n",
      "           1.25629371e-02,  9.13472287e-03,  1.67010706e-02],\n",
      "         [ 6.92509953e-03, -1.94874965e-03, -1.08433189e-02, ...,\n",
      "          -2.86199898e-02, -1.08568631e-02,  1.76268723e-03],\n",
      "         [-3.47630754e-02, -4.08676155e-02, -1.76608097e-02, ...,\n",
      "          -1.65180527e-02,  1.41871683e-02, -2.77192928e-02],\n",
      "         ...,\n",
      "         [-3.60747799e-05,  1.25048391e-04,  4.27067392e-02, ...,\n",
      "           3.03697102e-02,  5.26862126e-03, -7.75791751e-03],\n",
      "         [ 5.22814272e-03,  3.55891064e-02,  4.00468297e-02, ...,\n",
      "          -3.91718931e-02,  1.53721254e-02, -2.80213393e-02],\n",
      "         [ 8.22906662e-03, -2.09063347e-02,  4.12246101e-02, ...,\n",
      "           5.41590201e-03, -1.94895994e-02,  1.33307083e-02]],\n",
      "\n",
      "        [[-1.75995547e-02,  1.33226747e-02,  4.79265116e-03, ...,\n",
      "           3.08144297e-02,  5.92808425e-03,  2.75909416e-02],\n",
      "         [-2.12287363e-02,  3.73519920e-02,  5.06441575e-03, ...,\n",
      "           3.25928815e-02, -1.45289013e-02, -3.63701396e-02],\n",
      "         [ 1.21062612e-02, -1.72464270e-02, -5.04742516e-03, ...,\n",
      "           1.33419782e-02, -3.11648902e-02,  1.13765448e-02],\n",
      "         ...,\n",
      "         [ 2.60578673e-02,  3.50068770e-02,  3.71943414e-02, ...,\n",
      "           5.93282795e-03,  1.63202651e-03, -3.10969353e-02],\n",
      "         [-2.11694613e-02,  1.39735555e-02,  1.87783353e-02, ...,\n",
      "           1.84620265e-02, -6.02453062e-03,  3.86537649e-02],\n",
      "         [ 7.61347963e-03, -1.36647103e-02, -3.52959000e-02, ...,\n",
      "          -3.89531963e-02,  5.56456065e-03, -1.41226612e-02]],\n",
      "\n",
      "        [[-3.38194631e-02,  2.84826607e-02,  3.85427624e-02, ...,\n",
      "           1.68548506e-02,  2.76945159e-02,  2.33616885e-02],\n",
      "         [ 3.61935385e-02, -2.88651623e-02,  4.16025929e-02, ...,\n",
      "           2.93536335e-02, -1.37824211e-02,  2.74867080e-02],\n",
      "         [ 3.31576206e-02, -3.89800332e-02,  1.05086237e-03, ...,\n",
      "           2.03542691e-02,  1.60979442e-02, -2.98257377e-02],\n",
      "         ...,\n",
      "         [-2.84691574e-04,  2.86377873e-02,  3.70614901e-02, ...,\n",
      "           1.50839956e-02,  2.32211798e-02, -1.13844732e-02],\n",
      "         [-2.25482099e-02, -7.20875850e-03, -7.81299267e-03, ...,\n",
      "          -1.68694165e-02, -3.27695459e-02,  1.85703989e-02],\n",
      "         [ 2.79189162e-02,  1.35749048e-02,  4.88572661e-03, ...,\n",
      "          -2.05275733e-02, -4.13489379e-02, -1.12026967e-02]]],\n",
      "\n",
      "\n",
      "       [[[ 5.85267693e-03, -3.10657043e-02,  2.99593508e-02, ...,\n",
      "           1.80567503e-02,  9.59485769e-03, -2.12445669e-02],\n",
      "         [ 2.55949013e-02, -3.78607251e-02, -3.10788825e-02, ...,\n",
      "          -1.36395926e-02,  2.91798562e-02,  2.61008795e-02],\n",
      "         [ 3.88025977e-02, -1.07262731e-02,  3.46677117e-02, ...,\n",
      "          -3.25431786e-02, -2.97423378e-02,  2.28629040e-04],\n",
      "         ...,\n",
      "         [ 2.30423491e-02,  3.14652324e-02,  2.87560020e-02, ...,\n",
      "          -1.25040533e-02,  8.20346083e-03, -9.97276232e-03],\n",
      "         [ 1.61592215e-02, -1.28715876e-02, -3.68916690e-02, ...,\n",
      "           9.01370402e-03,  4.17498552e-04, -3.57346162e-02],\n",
      "         [-2.34480090e-02, -3.67690772e-02, -1.28474478e-02, ...,\n",
      "           2.97115799e-02,  2.25870386e-02,  2.09873188e-02]],\n",
      "\n",
      "        [[ 2.00278237e-02,  2.22435519e-02, -5.90521470e-03, ...,\n",
      "          -1.38483942e-02, -3.70222889e-02, -4.05819863e-02],\n",
      "         [-6.34367391e-03, -1.56043405e-02, -2.32879184e-02, ...,\n",
      "          -3.07890791e-02,  4.01785709e-02,  1.19034748e-03],\n",
      "         [ 1.40705612e-02, -1.98639743e-02, -1.70298647e-02, ...,\n",
      "           4.48238279e-04,  2.13942546e-02, -8.89658171e-04],\n",
      "         ...,\n",
      "         [ 3.29257399e-02, -2.60201152e-02,  1.95154138e-02, ...,\n",
      "           1.41204912e-02,  2.34310655e-03, -1.07428320e-02],\n",
      "         [ 3.23904417e-02, -1.93689968e-02, -7.74615817e-03, ...,\n",
      "           3.76699753e-02, -4.63150395e-03, -2.30019167e-02],\n",
      "         [ 3.46950106e-02,  1.45221381e-02, -3.99865136e-02, ...,\n",
      "           3.58383991e-02,  1.59548335e-02,  1.16713988e-02]],\n",
      "\n",
      "        [[-3.84135395e-02,  3.59655768e-02,  3.67261209e-02, ...,\n",
      "           1.40726315e-02,  2.79685818e-02, -8.68437625e-03],\n",
      "         [ 4.06317934e-02,  3.26275118e-02, -2.05395315e-02, ...,\n",
      "          -1.74826831e-02,  3.13782878e-02,  8.18094797e-03],\n",
      "         [-1.87192224e-02,  1.74652375e-02, -1.03478390e-03, ...,\n",
      "          -2.84220502e-02,  9.19463672e-03, -2.70445235e-02],\n",
      "         ...,\n",
      "         [-3.37319970e-02,  5.10906754e-03,  4.80249804e-03, ...,\n",
      "           2.71781906e-02,  2.45734863e-02, -3.52872759e-02],\n",
      "         [-1.87903028e-02, -3.97544503e-02, -1.04965335e-02, ...,\n",
      "          -3.28743681e-02,  1.26893236e-03, -2.67717130e-02],\n",
      "         [ 2.95641087e-02, -4.12448347e-02,  2.66017076e-02, ...,\n",
      "           3.14094462e-02,  2.94472054e-02,  1.50814084e-02]]],\n",
      "\n",
      "\n",
      "       [[[ 5.68909431e-03, -1.89000629e-02, -3.57524492e-02, ...,\n",
      "          -1.37803312e-02, -3.83978263e-02,  3.95146869e-02],\n",
      "         [ 3.54534201e-02, -3.04417461e-02,  1.15164733e-02, ...,\n",
      "          -2.32190844e-02,  2.01745555e-02,  2.14424878e-02],\n",
      "         [-3.22369449e-02, -2.32383348e-02, -1.63759273e-02, ...,\n",
      "          -1.82048306e-02,  3.80834565e-02,  2.67510414e-02],\n",
      "         ...,\n",
      "         [-2.92894803e-02,  1.65734664e-02,  2.15279367e-02, ...,\n",
      "           1.56524368e-02, -8.15407187e-03,  2.45651584e-02],\n",
      "         [-2.49561947e-02, -1.08091384e-02,  2.58876216e-02, ...,\n",
      "          -3.06014763e-03,  2.98446771e-02, -1.03163542e-02],\n",
      "         [ 3.72358859e-02,  5.93189616e-03,  7.76502071e-04, ...,\n",
      "           3.74459922e-02, -5.82675124e-03, -9.91493557e-03]],\n",
      "\n",
      "        [[-8.90900905e-04, -2.93555297e-02, -4.46414808e-03, ...,\n",
      "           1.51897445e-02, -2.69751158e-02,  7.12922588e-03],\n",
      "         [ 3.09253428e-02, -1.17044467e-02,  4.71520051e-02, ...,\n",
      "           1.18037444e-02, -6.04391983e-03, -1.21820718e-03],\n",
      "         [ 2.03830414e-02, -2.63565443e-02, -4.72789165e-03, ...,\n",
      "          -2.37903353e-02,  1.18411649e-02,  2.70994250e-02],\n",
      "         ...,\n",
      "         [-1.50230592e-02,  7.98966177e-03,  3.06982659e-02, ...,\n",
      "           1.02950493e-02, -1.55433984e-02,  3.48223634e-02],\n",
      "         [-3.03928144e-02,  1.77267101e-02, -3.34844589e-02, ...,\n",
      "           4.11323532e-02, -6.15237188e-03,  3.30243818e-02],\n",
      "         [-2.53857430e-02, -2.46210694e-02, -3.41215287e-03, ...,\n",
      "           1.13514974e-03,  2.15030015e-02, -2.06366833e-02]],\n",
      "\n",
      "        [[-3.61999348e-02,  8.16201698e-03, -4.00158688e-02, ...,\n",
      "          -3.68790068e-02,  1.30162071e-02,  2.68620234e-02],\n",
      "         [-2.03712191e-03,  2.39845216e-02, -1.02239531e-02, ...,\n",
      "           2.64528301e-02, -3.61691974e-02, -1.04851415e-02],\n",
      "         [ 2.73413155e-02, -1.68717187e-02, -3.13692018e-02, ...,\n",
      "           4.01808582e-02, -2.88476087e-02,  5.41008823e-03],\n",
      "         ...,\n",
      "         [ 2.00128537e-02,  2.15569139e-03,  1.05833942e-02, ...,\n",
      "           3.80696431e-02,  2.99004689e-02, -2.80720610e-02],\n",
      "         [-1.49689857e-02, -3.17634866e-02, -3.21634673e-02, ...,\n",
      "          -2.47874148e-02,  7.02652615e-03, -8.88030976e-03],\n",
      "         [-3.60690579e-02, -1.38445264e-02, -1.59982871e-03, ...,\n",
      "           1.81184765e-02,  1.74645614e-02,  2.30861828e-03]]]],\n",
      "      dtype=float32), array([-3.98978265e-03, -1.18210271e-03,  1.42590916e-02,  6.92192093e-02,\n",
      "        3.89777916e-03,  1.12411734e-02,  4.02755886e-02, -5.39260823e-03,\n",
      "        1.51297813e-02,  1.07867941e-01,  1.08473981e-03,  6.86506554e-02,\n",
      "        1.81755779e-04,  1.75036713e-02,  1.36259675e-01,  1.02681201e-03,\n",
      "       -1.02867128e-03,  1.44265771e-01,  4.59231297e-03, -1.63597311e-03,\n",
      "        1.00775205e-01,  2.15841137e-04,  9.99350920e-02,  1.05178118e-01,\n",
      "        2.35586595e-02,  2.33748835e-02, -3.11569171e-03,  1.06511787e-01,\n",
      "        2.11558631e-03,  1.92813147e-02,  2.04147510e-02, -3.56189418e-03,\n",
      "       -1.09252974e-03,  1.39491614e-02,  1.19547741e-02,  2.59265155e-02,\n",
      "        5.44790365e-02,  6.25278044e-04, -1.15609926e-03, -3.59426462e-03,\n",
      "       -2.20540329e-03,  1.10383751e-02, -2.41046437e-04,  2.55820975e-02,\n",
      "       -3.02241952e-03,  1.95428152e-02,  6.34172931e-04,  3.18961218e-03,\n",
      "       -3.10969655e-03,  3.54988091e-02,  4.23664451e-02,  6.69118613e-02,\n",
      "        3.37292962e-02,  5.71906082e-02,  1.12463422e-01,  9.54734441e-03,\n",
      "       -2.80084577e-03, -2.02229549e-03,  6.38399273e-02,  9.92370769e-04,\n",
      "        9.50861722e-02,  2.13470887e-02,  5.93753345e-02, -4.05453284e-05,\n",
      "        2.01334842e-02,  8.66825506e-02,  5.14337942e-02,  5.83447143e-02,\n",
      "       -2.12161662e-03,  5.03883883e-02,  3.32901664e-02, -6.30343426e-03,\n",
      "        1.29094673e-03,  2.47880220e-02,  8.21575150e-02,  7.36462418e-03,\n",
      "        1.17705227e-03,  1.33270128e-02,  5.25416955e-02,  2.61897966e-02,\n",
      "       -2.28617154e-03,  1.49455881e-02,  9.64957289e-03, -1.45238810e-04,\n",
      "        7.30885938e-02,  2.15475280e-02,  8.92340764e-02,  1.19167365e-01,\n",
      "       -2.40576849e-03, -3.19711282e-04, -2.31734477e-03,  6.79456061e-05,\n",
      "        3.36951367e-03,  1.01889379e-01,  3.25440094e-02,  2.04456504e-02,\n",
      "        9.17361006e-02,  5.40974736e-02,  1.20174266e-01,  3.63655239e-02,\n",
      "        1.88770965e-02, -2.55413516e-03,  1.70778763e-02,  1.86926648e-01,\n",
      "        1.36093795e-03,  4.08998504e-02, -1.55585725e-03, -2.10415176e-03,\n",
      "        1.03627101e-01,  7.51798078e-02,  8.18840489e-02,  5.66950627e-02,\n",
      "        3.63710970e-02, -3.21722613e-03, -1.96366556e-04, -1.58298761e-03,\n",
      "       -2.51171901e-03,  7.25922137e-02,  9.47652459e-02,  3.26543930e-03,\n",
      "       -1.66498404e-03,  2.39065709e-03,  1.27950183e-03, -6.39357255e-04,\n",
      "        5.16517535e-02,  6.80060387e-02,  5.56999957e-03, -4.04840917e-04,\n",
      "        2.56517045e-02,  2.55322754e-02,  1.98681783e-02,  4.87248637e-02,\n",
      "        1.38888136e-02,  5.00791930e-02,  2.56158449e-02,  1.49776757e-01,\n",
      "       -3.49415047e-03,  2.78564263e-02,  1.42084926e-01, -5.27214585e-03,\n",
      "       -4.09053871e-03, -1.06260832e-03,  7.97221586e-02,  1.05152115e-01,\n",
      "       -4.16776165e-03,  6.20213291e-03,  2.94758566e-03,  7.83996731e-02,\n",
      "        2.95217931e-02,  2.14111041e-02,  3.21684568e-03,  8.80727246e-02,\n",
      "        5.55450171e-02, -9.96686285e-05,  3.25266346e-02, -5.92071936e-03,\n",
      "        2.27313004e-02, -1.25607825e-03,  1.52355833e-02,  7.96201080e-02,\n",
      "        1.78155806e-02, -1.86792959e-03,  3.42356344e-03,  4.31804210e-02,\n",
      "        7.45269982e-03, -2.05700705e-03,  1.03552304e-01, -1.95730152e-03,\n",
      "        6.10942431e-02,  1.11380860e-01,  7.49946712e-03, -5.46576455e-04,\n",
      "        5.11536971e-02,  8.35127831e-02,  6.38561994e-02, -4.25230501e-05,\n",
      "        1.13353111e-01,  3.30799073e-02,  1.07201422e-02, -9.90843750e-04,\n",
      "        7.34251067e-02,  6.20375015e-02,  2.59448425e-03,  5.92815280e-02,\n",
      "        8.50612763e-04,  4.13954034e-02,  1.43167982e-02,  5.14480239e-03,\n",
      "       -3.30172875e-03,  1.18992761e-01, -9.27356072e-03,  3.70769762e-02,\n",
      "        4.07044366e-02, -9.42925224e-04, -2.30940455e-03,  1.03443824e-01,\n",
      "        6.63901097e-04, -1.21019792e-03,  6.00214563e-02, -2.40015262e-03,\n",
      "        5.76203689e-02,  1.00516416e-02,  8.56103897e-02,  7.02047423e-02,\n",
      "        3.99240293e-03,  4.42009382e-02,  6.21556258e-03, -1.56974664e-03,\n",
      "        2.80651785e-02,  8.28364771e-03, -8.99279141e-04,  2.85326573e-03,\n",
      "        1.90207828e-02, -6.14714157e-03, -3.63530824e-04, -5.28797973e-03,\n",
      "       -1.30663381e-03,  2.03596223e-02,  5.99536039e-02,  1.73330292e-04,\n",
      "        3.13206390e-03,  5.73118068e-02,  9.58047155e-03,  3.09475255e-03,\n",
      "        8.37173611e-02,  6.18163832e-02,  2.26796214e-02,  3.51108005e-03,\n",
      "        3.19277234e-02,  3.05005666e-02,  9.03295353e-03,  1.19216807e-01,\n",
      "       -6.64719963e-04,  4.60846610e-02,  1.95304137e-02,  3.95090290e-04,\n",
      "        2.85532861e-03, -1.95022312e-03,  3.84242460e-02,  7.14084483e-04,\n",
      "        7.06304088e-02,  1.33956270e-02,  4.36668359e-02,  4.07753997e-02,\n",
      "        3.77747905e-03,  3.48434970e-02,  4.09998326e-03,  7.88275339e-03,\n",
      "       -4.89290571e-03,  6.02914812e-03,  6.68988600e-02,  4.47232686e-02,\n",
      "        6.74702898e-02,  8.38266220e-03, -1.13441946e-03,  9.27260611e-03],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'padding': 'valid', 'data_format': 'channels_first', 'strides': (2, 2), 'pool_size': (2, 2), 'name': 'max_pooling2d_2', 'trainable': True} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'rate': 0.2, 'name': 'dropout_2', 'seed': None, 'trainable': True, 'noise_shape': None} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_size': (3, 3), 'name': 'conv2d_5', 'bias_regularizer': None, 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'filters': 256, 'use_bias': True, 'padding': 'same', 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[[[-2.92526614e-02, -2.74499767e-02, -2.50900239e-02, ...,\n",
      "          -2.85964496e-02,  2.07122099e-02, -6.35244465e-03],\n",
      "         [ 1.15605704e-02, -3.53721716e-02, -3.54633443e-02, ...,\n",
      "           7.73897453e-04, -2.42368877e-02, -1.42848408e-02],\n",
      "         [ 2.70632952e-02, -9.09330137e-03,  3.46890390e-02, ...,\n",
      "          -7.07281660e-03,  1.26382336e-02,  4.08359021e-02],\n",
      "         ...,\n",
      "         [ 3.62348221e-02,  1.32539188e-02,  3.32424566e-02, ...,\n",
      "          -1.66336242e-02, -3.30430977e-02,  2.55743414e-02],\n",
      "         [-2.04842389e-02,  2.43756715e-02, -1.51633453e-02, ...,\n",
      "           2.51220223e-02, -2.66553555e-02, -1.74114686e-02],\n",
      "         [ 1.07122427e-02, -1.33838458e-02,  4.76788031e-03, ...,\n",
      "           2.83889715e-02, -7.12309359e-03, -5.89415757e-03]],\n",
      "\n",
      "        [[-1.50684016e-02,  1.41358711e-02,  2.97889998e-03, ...,\n",
      "           3.23733203e-02, -3.52498218e-02,  1.67883988e-02],\n",
      "         [ 1.97152812e-02, -2.35651806e-02, -2.96043083e-02, ...,\n",
      "           2.31159609e-02, -6.97796606e-03,  9.83241014e-03],\n",
      "         [ 3.62345413e-03,  1.61305852e-02,  4.00411300e-02, ...,\n",
      "           3.17591839e-02,  2.50956696e-02, -2.80327047e-03],\n",
      "         ...,\n",
      "         [ 2.08592005e-02,  7.49498140e-03,  3.36707830e-02, ...,\n",
      "          -5.38862217e-03, -1.13937045e-02,  2.15238668e-02],\n",
      "         [-1.19784018e-02, -1.20163951e-02,  4.92312200e-03, ...,\n",
      "          -2.47806162e-02,  2.34750137e-02, -2.48365086e-02],\n",
      "         [ 3.56195979e-02,  1.57268420e-02, -2.86832061e-02, ...,\n",
      "          -2.79487483e-02, -1.30066564e-02, -2.48526353e-02]],\n",
      "\n",
      "        [[-2.23095231e-02,  3.18944571e-03,  2.63375193e-02, ...,\n",
      "           6.35436270e-03, -1.68381594e-02,  2.36525759e-02],\n",
      "         [ 5.21755684e-03, -2.88448054e-02,  2.22331341e-02, ...,\n",
      "           2.19445676e-02, -4.87768417e-03, -3.58221866e-02],\n",
      "         [-4.58627148e-03, -1.38309682e-02, -1.79125853e-02, ...,\n",
      "           1.42595293e-02,  2.85110865e-02,  2.77227834e-02],\n",
      "         ...,\n",
      "         [ 3.71730654e-03,  2.99032815e-02, -5.91064291e-03, ...,\n",
      "          -3.48700583e-02, -2.89695058e-02,  1.04884673e-02],\n",
      "         [-2.73986161e-03,  1.86428539e-02, -9.00059513e-06, ...,\n",
      "           2.88752988e-02, -2.79885828e-02, -1.30672734e-02],\n",
      "         [ 1.28624747e-02,  2.73810998e-02,  8.53124820e-03, ...,\n",
      "          -7.71983853e-03, -1.48351314e-02, -3.05064116e-02]]],\n",
      "\n",
      "\n",
      "       [[[-2.56383466e-03,  1.70488562e-02,  5.44159964e-04, ...,\n",
      "          -2.25264858e-02, -3.13245729e-02, -2.90681911e-03],\n",
      "         [ 3.53931710e-02, -2.22188570e-02,  3.35472487e-02, ...,\n",
      "           8.43572617e-03, -1.67771783e-02, -2.04817345e-03],\n",
      "         [-1.08525419e-04,  1.29361311e-03, -1.33980745e-02, ...,\n",
      "           1.38001610e-02,  3.31242420e-02, -1.88463368e-02],\n",
      "         ...,\n",
      "         [-1.94724221e-02,  4.30863118e-03,  3.45456526e-02, ...,\n",
      "           2.54610665e-02, -1.72751751e-02,  2.32070833e-02],\n",
      "         [ 6.89838594e-03,  7.78438244e-03, -7.24508893e-03, ...,\n",
      "          -2.25272719e-02,  1.07652349e-02, -6.31193118e-03],\n",
      "         [ 2.59705540e-02, -1.48488916e-02, -3.25054601e-02, ...,\n",
      "          -2.49159727e-02, -2.66836048e-03,  9.53299925e-03]],\n",
      "\n",
      "        [[-1.01609053e-02, -2.74647996e-02,  1.84292141e-02, ...,\n",
      "          -3.24731991e-02,  2.71446574e-02, -1.68275796e-02],\n",
      "         [ 1.38812196e-02,  3.56341042e-02,  7.53806555e-04, ...,\n",
      "          -1.91958901e-02,  1.85582396e-02, -3.58893373e-03],\n",
      "         [-2.84496392e-03, -1.92917958e-02,  3.93770784e-02, ...,\n",
      "           3.10572404e-02, -1.37815746e-02, -8.75820965e-03],\n",
      "         ...,\n",
      "         [-4.19459213e-03, -1.90791655e-02, -1.20597770e-02, ...,\n",
      "          -3.16171981e-02, -2.88974512e-02,  9.08832159e-03],\n",
      "         [-1.57638863e-02,  1.92501042e-02,  2.20145322e-02, ...,\n",
      "           2.11766232e-02,  5.02218492e-03, -1.87202375e-02],\n",
      "         [-2.07621306e-02,  3.60131077e-02,  9.24869347e-03, ...,\n",
      "          -2.99260542e-02, -2.96787545e-02,  1.90655293e-03]],\n",
      "\n",
      "        [[-3.26157548e-02,  4.99257166e-03, -2.55026966e-02, ...,\n",
      "          -3.10396380e-03,  3.60088013e-02,  4.99983924e-03],\n",
      "         [-2.93962155e-02, -3.19476500e-02, -3.15387286e-02, ...,\n",
      "           3.60611007e-02,  5.60758449e-03, -1.51541252e-02],\n",
      "         [-5.27300267e-03, -2.57061869e-02, -2.31424137e-03, ...,\n",
      "          -2.06351336e-02, -2.35098414e-02,  1.63608938e-02],\n",
      "         ...,\n",
      "         [ 1.67543311e-02,  5.38098440e-03, -2.89921626e-03, ...,\n",
      "          -3.41435932e-02,  2.19883714e-02,  7.50074442e-03],\n",
      "         [-3.30014005e-02,  1.40410196e-02, -3.25886346e-02, ...,\n",
      "           3.07530332e-02,  1.14389192e-02, -1.05989529e-02],\n",
      "         [-3.49929295e-02,  4.53576213e-03, -5.31662861e-03, ...,\n",
      "          -2.43469588e-02, -3.84609884e-04, -2.17419732e-02]]],\n",
      "\n",
      "\n",
      "       [[[ 1.62660405e-02,  2.27199700e-02,  2.11636852e-02, ...,\n",
      "          -2.45119464e-02,  7.66993081e-03, -1.55666005e-02],\n",
      "         [-7.27998186e-03, -2.76843477e-02,  8.73439200e-03, ...,\n",
      "          -1.12780202e-02, -2.94622909e-02,  1.16832145e-02],\n",
      "         [-7.26466812e-03,  2.05734875e-02, -7.38060661e-03, ...,\n",
      "           1.98901687e-02, -3.19489613e-02, -2.46234015e-02],\n",
      "         ...,\n",
      "         [-2.96646059e-02,  2.96499506e-02,  7.13736843e-03, ...,\n",
      "          -5.49650146e-03, -1.13650719e-02, -9.38201789e-03],\n",
      "         [ 2.26947498e-02, -2.52982378e-02, -5.13915811e-03, ...,\n",
      "           1.88861135e-02, -1.10565601e-02,  3.10490094e-02],\n",
      "         [ 3.29646654e-02,  4.98608453e-03, -7.95511622e-03, ...,\n",
      "           1.87984649e-02,  8.14924669e-03, -4.08297731e-03]],\n",
      "\n",
      "        [[ 2.68808883e-02,  3.14715281e-02, -1.81649234e-02, ...,\n",
      "          -2.29947288e-02, -3.29991616e-02,  2.69785598e-02],\n",
      "         [-1.87101699e-02,  2.41316650e-02, -1.74267367e-02, ...,\n",
      "          -1.29195405e-02,  2.02596784e-02,  2.63161454e-02],\n",
      "         [-4.14850801e-04,  1.43149951e-02,  5.31682791e-03, ...,\n",
      "          -1.61502771e-02, -1.55316032e-02, -3.01942099e-02],\n",
      "         ...,\n",
      "         [-2.79950965e-02, -3.18722427e-02, -2.20582671e-02, ...,\n",
      "           6.93252450e-03,  1.36474892e-03, -3.28096263e-02],\n",
      "         [ 3.23772617e-02, -2.75620539e-02, -1.73891010e-03, ...,\n",
      "          -2.55479217e-02,  1.64823991e-03, -2.60605030e-02],\n",
      "         [ 2.89614853e-02,  1.33423992e-02, -3.25879529e-02, ...,\n",
      "           9.68449749e-03, -1.54429460e-02, -1.26817299e-03]],\n",
      "\n",
      "        [[ 7.46752368e-03, -1.66946538e-02,  8.81380402e-03, ...,\n",
      "           4.33246186e-03,  2.82790083e-02,  1.28362491e-03],\n",
      "         [-1.86784696e-02,  4.47085546e-03,  3.80611932e-03, ...,\n",
      "          -6.81525888e-03,  1.39445709e-02, -1.18978200e-02],\n",
      "         [ 1.03545636e-02,  2.17865054e-02,  1.85798071e-02, ...,\n",
      "          -8.51808768e-03,  2.14107446e-02,  5.70389815e-03],\n",
      "         ...,\n",
      "         [ 5.46016172e-03,  2.79946383e-02,  8.34631640e-03, ...,\n",
      "          -2.64810212e-03, -5.27443655e-04,  3.02642509e-02],\n",
      "         [-2.57398579e-02,  1.28066470e-03, -1.69681199e-02, ...,\n",
      "          -1.74566805e-02,  2.90409904e-02,  2.21379641e-02],\n",
      "         [-3.06221955e-02, -1.54045662e-02,  2.44610179e-02, ...,\n",
      "           2.74881870e-02,  1.09434277e-02, -8.26795958e-03]]]],\n",
      "      dtype=float32), array([ 1.26009888e-03,  2.51207501e-02,  9.82283056e-03,  1.48346033e-02,\n",
      "        3.37221939e-03,  2.23263167e-02,  7.17792148e-03, -6.54564193e-03,\n",
      "        1.08091468e-02, -5.74837578e-03,  7.85982236e-03,  1.25618745e-03,\n",
      "        2.58702375e-02,  7.41698779e-03,  1.40000571e-04, -4.89446754e-03,\n",
      "        1.11518428e-03,  4.02002782e-03, -4.12797890e-05,  1.99526772e-02,\n",
      "        2.66097821e-02, -2.40885047e-03,  1.20013477e-02,  1.17240390e-02,\n",
      "        5.76044759e-03,  2.41811238e-02,  7.97663629e-03,  7.08678551e-03,\n",
      "        7.83209410e-03,  9.34397709e-03,  8.59220140e-03,  1.86112877e-02,\n",
      "        6.54308498e-03,  3.06778755e-02, -6.77236542e-03, -5.12412703e-03,\n",
      "        7.77044240e-03, -5.16040204e-03, -4.50286345e-04,  2.22395975e-02,\n",
      "        5.96206356e-03,  2.70500779e-02,  5.55616710e-03, -1.19374658e-03,\n",
      "        1.05920713e-02,  7.88009318e-04,  2.12264601e-02, -8.60590395e-03,\n",
      "       -1.88071060e-03, -1.28021352e-02,  2.54794984e-04,  1.51358517e-02,\n",
      "        1.47673730e-02, -4.33799950e-03,  1.96451992e-02, -2.24314281e-03,\n",
      "        6.36293320e-04,  6.25354151e-05, -3.15465336e-03,  3.58097404e-02,\n",
      "        1.95165141e-03, -1.77150383e-03,  2.39002955e-04,  5.71403466e-03,\n",
      "       -2.11501401e-03, -4.95995861e-03,  3.65137006e-03, -1.34485064e-03,\n",
      "        1.15272757e-02, -3.63652158e-04,  3.93991405e-03, -4.80764499e-03,\n",
      "        3.15852761e-02, -1.09044032e-03,  9.09301452e-05,  8.42481561e-04,\n",
      "        6.76957518e-03, -3.31901480e-03,  2.83302274e-03, -2.04309402e-03,\n",
      "        1.21464091e-03,  3.24506895e-03,  2.22319784e-03, -3.43783596e-03,\n",
      "        5.61625743e-03,  3.63715179e-03,  1.69374570e-02,  7.23606069e-03,\n",
      "        1.37872971e-03, -6.33649994e-04,  1.63043405e-05,  1.26682632e-02,\n",
      "        1.56341288e-02,  1.36045879e-03,  2.79286653e-02,  3.11558507e-02,\n",
      "       -5.93189452e-06, -6.86916476e-03,  3.95782627e-02, -2.99808453e-03,\n",
      "        7.90259615e-03,  4.80901524e-02,  2.61753425e-02, -4.85089933e-03,\n",
      "        4.86587407e-03,  3.41081596e-03, -3.80770507e-04, -2.33565341e-04,\n",
      "        4.49690409e-03, -1.93108141e-03,  6.06507948e-03, -7.14544149e-05,\n",
      "       -4.29249427e-04, -3.42516229e-04,  2.47930419e-02,  2.38571577e-02,\n",
      "       -4.04009642e-03,  1.38759604e-02, -9.84236877e-03,  4.71650856e-03,\n",
      "        4.11409093e-03, -5.70113165e-03,  4.04726788e-02, -2.87705846e-03,\n",
      "        1.50630493e-02,  1.94118416e-03,  3.75213730e-03,  2.39629205e-02,\n",
      "        8.42463132e-03, -2.47961050e-03,  2.90391780e-03,  1.10667888e-02,\n",
      "       -1.42354146e-03,  2.59794178e-03,  9.10546165e-03, -8.93738412e-04,\n",
      "        7.22957123e-03, -5.27329929e-03,  3.34325782e-03, -1.68189744e-03,\n",
      "        1.25111882e-02,  1.22809573e-03,  2.69195251e-03,  7.72921368e-04,\n",
      "       -2.85185874e-03, -2.44360697e-03, -3.52847320e-03,  4.98788152e-03,\n",
      "        5.24338474e-03,  3.96419987e-02, -6.96671987e-03,  1.44893490e-02,\n",
      "       -3.39980936e-03,  2.18008496e-02,  1.40387621e-02,  4.89518931e-03,\n",
      "       -1.70290796e-03, -7.97149623e-05,  3.26615721e-02,  3.48555446e-02,\n",
      "        4.73823026e-03,  1.89316534e-02,  4.88151237e-03, -6.84516388e-04,\n",
      "       -3.47462320e-03, -2.85046524e-03,  8.41332367e-05,  2.98565906e-03,\n",
      "       -1.74797722e-03,  2.01410451e-03,  1.40397698e-02,  6.02543028e-03,\n",
      "        5.02198981e-03, -7.12915789e-04,  5.78760373e-05,  1.09840563e-04,\n",
      "        4.73206863e-03, -4.37307550e-04,  4.62479144e-03,  8.47456604e-03,\n",
      "        1.96780544e-03, -8.82490911e-03,  1.08086597e-03,  1.05871735e-02,\n",
      "        3.51491850e-04,  5.60902916e-02,  3.59845534e-03,  7.02568504e-04,\n",
      "        1.38735631e-02,  1.31772021e-02,  1.56652145e-02,  1.35591375e-02,\n",
      "        1.96049828e-03, -5.96503145e-04,  3.16287056e-02,  1.11129656e-02,\n",
      "        2.14403197e-02,  1.55381141e-02,  1.10073853e-02, -5.79424528e-03,\n",
      "        1.17980107e-03,  3.27151292e-03, -8.27371329e-03,  1.11160884e-02,\n",
      "       -2.32841959e-03, -5.56727219e-03,  1.27769541e-03,  1.05897514e-02,\n",
      "        5.89236524e-03,  3.19314525e-02,  1.18324673e-02,  7.38477940e-03,\n",
      "        1.59242600e-02,  1.63789578e-02,  1.77715831e-02, -1.62699696e-04,\n",
      "       -6.75995508e-03,  6.87625492e-03,  2.96673970e-03,  6.08398858e-03,\n",
      "        2.45288219e-02,  3.73264495e-03,  2.65788082e-02, -4.58925590e-03,\n",
      "        1.27694998e-02,  8.92154523e-04, -6.44749147e-04,  4.23461944e-03,\n",
      "        2.61696870e-03,  5.81391272e-04, -7.52010988e-03,  1.37014054e-02,\n",
      "        1.58523617e-04, -1.85809535e-04,  1.33225713e-02,  1.14905117e-02,\n",
      "        2.44203926e-04, -6.78433012e-03, -2.45842338e-03,  1.61096361e-03,\n",
      "        6.98517018e-04,  9.63450223e-03,  1.90024935e-02,  2.91919522e-03,\n",
      "        2.06765886e-02, -1.76292576e-03,  6.29770453e-04,  6.58024428e-03,\n",
      "        5.30677242e-03,  3.94927152e-03, -7.63299828e-03,  1.63359265e-03,\n",
      "       -6.01828157e-04, -2.69279262e-04, -8.93526652e-04,  1.92516353e-02],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_size': (3, 3), 'name': 'conv2d_6', 'bias_regularizer': None, 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'filters': 256, 'use_bias': True, 'padding': 'valid', 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[[[ 0.03122446, -0.02781327, -0.02894998, ..., -0.03500555,\n",
      "           0.02199835, -0.02526066],\n",
      "         [ 0.01267773, -0.0109235 ,  0.02798333, ..., -0.01694916,\n",
      "           0.01393856,  0.01285979],\n",
      "         [ 0.02818282, -0.0178725 , -0.03507611, ..., -0.00837052,\n",
      "          -0.02774097,  0.01604588],\n",
      "         ...,\n",
      "         [-0.02046458, -0.01791687,  0.00349032, ..., -0.02914568,\n",
      "           0.02285402, -0.01581593],\n",
      "         [-0.00192385, -0.02460909, -0.01847829, ..., -0.03358748,\n",
      "           0.02460573, -0.0155954 ],\n",
      "         [ 0.01219428,  0.03375032,  0.00011298, ...,  0.01698548,\n",
      "          -0.00821977, -0.01804131]],\n",
      "\n",
      "        [[-0.01604264, -0.02020148,  0.0300421 , ..., -0.02970299,\n",
      "          -0.02320589,  0.03760651],\n",
      "         [ 0.01105908,  0.02681559,  0.00592989, ..., -0.0063922 ,\n",
      "          -0.02066022,  0.01356243],\n",
      "         [-0.01271735,  0.01502105, -0.00189997, ..., -0.02842471,\n",
      "          -0.03586023,  0.04564926],\n",
      "         ...,\n",
      "         [ 0.01647933,  0.01363806, -0.01270516, ...,  0.02897937,\n",
      "          -0.03070936,  0.02811055],\n",
      "         [-0.01326195,  0.02551403, -0.01860034, ...,  0.02003726,\n",
      "          -0.01340212,  0.0267854 ],\n",
      "         [-0.00118064, -0.022333  ,  0.02072218, ...,  0.00161857,\n",
      "           0.01514418, -0.02169704]],\n",
      "\n",
      "        [[ 0.02807547,  0.00304818,  0.00990987, ..., -0.03102503,\n",
      "          -0.01703453, -0.01111895],\n",
      "         [-0.01834811,  0.03855579, -0.03400086, ..., -0.00680531,\n",
      "          -0.01372589,  0.03930084],\n",
      "         [ 0.04234033, -0.01166368, -0.01095944, ...,  0.0231318 ,\n",
      "          -0.00017874,  0.02806031],\n",
      "         ...,\n",
      "         [-0.0288086 , -0.00649207, -0.00847208, ..., -0.03417912,\n",
      "          -0.00343297, -0.02267175],\n",
      "         [ 0.02614384,  0.00626853,  0.00107757, ..., -0.03264239,\n",
      "          -0.01368521, -0.0033059 ],\n",
      "         [-0.0115117 ,  0.00829065,  0.01102014, ..., -0.00671803,\n",
      "          -0.02462296, -0.00498525]]],\n",
      "\n",
      "\n",
      "       [[[ 0.00017645, -0.01607633,  0.03555825, ..., -0.01608294,\n",
      "           0.03424211, -0.01053569],\n",
      "         [ 0.00598397, -0.01798928, -0.0318426 , ..., -0.01079085,\n",
      "           0.02752396,  0.01282705],\n",
      "         [ 0.03519649,  0.03282692, -0.02973415, ...,  0.03553751,\n",
      "           0.00451427,  0.02464559],\n",
      "         ...,\n",
      "         [ 0.0342522 ,  0.01937687,  0.00656049, ..., -0.01939227,\n",
      "           0.01131277, -0.03254686],\n",
      "         [ 0.02887203,  0.0206668 ,  0.0205969 , ...,  0.00718635,\n",
      "          -0.01790149, -0.00406515],\n",
      "         [ 0.00213405,  0.03575297,  0.01350622, ..., -0.01692853,\n",
      "           0.01610349,  0.009401  ]],\n",
      "\n",
      "        [[-0.03099467,  0.02744923, -0.01699075, ..., -0.03180071,\n",
      "           0.03262966,  0.01119159],\n",
      "         [ 0.04877319, -0.02106068, -0.03119805, ..., -0.02952185,\n",
      "          -0.0334979 ,  0.01459354],\n",
      "         [ 0.05852138,  0.00106873, -0.01227332, ..., -0.02139951,\n",
      "           0.03163002,  0.02014356],\n",
      "         ...,\n",
      "         [-0.00806517,  0.009199  , -0.02933052, ..., -0.01840643,\n",
      "           0.00624188,  0.0248063 ],\n",
      "         [-0.01772383, -0.01691049,  0.01500687, ..., -0.01196329,\n",
      "          -0.03467287,  0.03288516],\n",
      "         [ 0.00022369,  0.0187951 , -0.01676953, ..., -0.0023994 ,\n",
      "           0.02992455,  0.02140785]],\n",
      "\n",
      "        [[-0.01966543,  0.00568323,  0.00543247, ..., -0.00159961,\n",
      "           0.00446635,  0.03038563],\n",
      "         [ 0.02449355,  0.03143593, -0.01696101, ..., -0.02972619,\n",
      "           0.01166245,  0.00266346],\n",
      "         [ 0.02062772,  0.0080312 , -0.00125484, ..., -0.01023828,\n",
      "           0.02637703, -0.01592566],\n",
      "         ...,\n",
      "         [-0.01026267, -0.00587189, -0.02274034, ...,  0.00492478,\n",
      "           0.0143357 , -0.02666245],\n",
      "         [ 0.00749328,  0.03145974,  0.01664227, ...,  0.0358249 ,\n",
      "           0.0020968 , -0.02682154],\n",
      "         [-0.00524709,  0.00313679,  0.00213793, ..., -0.0182112 ,\n",
      "          -0.0314292 ,  0.02976923]]],\n",
      "\n",
      "\n",
      "       [[[ 0.01041589,  0.03518323, -0.00978857, ..., -0.027847  ,\n",
      "           0.02158063,  0.01983402],\n",
      "         [ 0.04352633,  0.02362321, -0.00356228, ..., -0.02192845,\n",
      "          -0.01886894,  0.04524276],\n",
      "         [ 0.02726686,  0.00313351, -0.02446184, ..., -0.00143338,\n",
      "          -0.01983645,  0.04003459],\n",
      "         ...,\n",
      "         [ 0.02206485, -0.00175592,  0.03252452, ...,  0.00391898,\n",
      "          -0.02678566, -0.00758249],\n",
      "         [ 0.0095736 , -0.02394253,  0.00191723, ..., -0.00706764,\n",
      "          -0.00683489, -0.02533901],\n",
      "         [ 0.02688579,  0.02960801, -0.00122988, ..., -0.0134707 ,\n",
      "          -0.02814175,  0.00824767]],\n",
      "\n",
      "        [[ 0.02398867,  0.01028658, -0.01352706, ...,  0.01977144,\n",
      "           0.01374315,  0.01980796],\n",
      "         [-0.01721649,  0.0343755 , -0.01566116, ..., -0.02971944,\n",
      "           0.01263952,  0.00951326],\n",
      "         [-0.00760245, -0.00977687,  0.00346634, ..., -0.00157819,\n",
      "           0.02581352,  0.04415099],\n",
      "         ...,\n",
      "         [ 0.00916145,  0.00983519,  0.02037066, ..., -0.01499548,\n",
      "           0.01639533, -0.0237349 ],\n",
      "         [ 0.03126484, -0.03275011, -0.02052552, ..., -0.02931266,\n",
      "          -0.01818855, -0.01948376],\n",
      "         [-0.0034958 ,  0.02694081,  0.00423274, ..., -0.03195479,\n",
      "          -0.0308474 ,  0.03826758]],\n",
      "\n",
      "        [[ 0.00789271, -0.03411273,  0.01696725, ...,  0.00320707,\n",
      "          -0.0112115 , -0.00709392],\n",
      "         [-0.00495647,  0.03728157,  0.02951039, ..., -0.01628209,\n",
      "           0.0003521 ,  0.02238877],\n",
      "         [-0.00577196, -0.00994952, -0.01874207, ..., -0.0172834 ,\n",
      "          -0.02487567,  0.00012408],\n",
      "         ...,\n",
      "         [-0.00708555,  0.00690761,  0.03556256, ..., -0.03480485,\n",
      "           0.01593582,  0.00828721],\n",
      "         [ 0.01827196, -0.00021524,  0.02952977, ...,  0.00318524,\n",
      "          -0.02506404,  0.00354298],\n",
      "         [ 0.00245505, -0.01292966, -0.02845859, ...,  0.01116228,\n",
      "          -0.01079517, -0.00999498]]]], dtype=float32), array([-1.00312075e-02, -5.12103876e-03, -7.32382294e-03, -7.14158377e-05,\n",
      "        5.68973552e-03,  8.95940512e-03,  2.27632709e-02, -3.40813131e-04,\n",
      "       -9.52337217e-03, -2.17167311e-03,  3.60445515e-03,  3.53373191e-03,\n",
      "        2.59504933e-02,  2.62710340e-02,  3.45521816e-03,  2.00249851e-02,\n",
      "        2.04091286e-03, -8.09876714e-03,  6.03013963e-04,  1.07499969e-03,\n",
      "       -1.64095807e-04, -8.67013354e-03, -7.98200164e-03, -9.34728596e-04,\n",
      "       -1.31964160e-04, -7.87963683e-04, -1.85681414e-02, -5.21483133e-03,\n",
      "       -9.00903530e-03,  9.71442263e-04, -1.50959357e-03,  4.04329002e-02,\n",
      "        2.55207829e-02,  4.29108879e-03, -7.45440181e-03, -5.09805046e-04,\n",
      "       -1.05168829e-02, -5.73727628e-03, -8.28045327e-03, -9.01697390e-03,\n",
      "       -2.06110515e-02, -8.53989914e-05,  1.16544198e-02, -6.44862233e-03,\n",
      "       -3.25622829e-03, -4.72496171e-03, -8.43266398e-03, -2.16642884e-03,\n",
      "        1.57051235e-02, -7.31463451e-03,  3.07413004e-02, -4.39343305e-04,\n",
      "        1.43448096e-02,  2.83530820e-02,  1.80980843e-02, -2.25137379e-02,\n",
      "       -1.80892355e-03,  1.81625113e-02, -1.26306526e-02, -3.30009148e-03,\n",
      "        5.16661024e-03, -1.46745122e-03, -1.57135390e-02,  5.95140271e-03,\n",
      "        6.78139478e-02,  3.98489647e-03, -1.21108606e-03, -1.95541885e-03,\n",
      "       -1.27458567e-04, -5.30060148e-03,  1.47444531e-02,  2.29852349e-02,\n",
      "        1.29377237e-03,  4.67801699e-04, -4.42010956e-03,  1.90343428e-02,\n",
      "       -3.62431711e-05,  2.16714311e-02, -1.10933278e-02, -6.09388435e-03,\n",
      "       -7.58103980e-03, -4.24415199e-03, -1.24747946e-03, -6.51819806e-04,\n",
      "        2.47380114e-03, -3.98696902e-05,  1.50341550e-02, -1.43285654e-02,\n",
      "        2.76529249e-02,  1.49438269e-02,  1.33995721e-02,  2.40316633e-02,\n",
      "        9.39642545e-03,  3.18901800e-03,  3.48819979e-02,  5.12817642e-03,\n",
      "        7.34817877e-05,  1.40491512e-03,  2.85695326e-02,  1.62398443e-02,\n",
      "        2.23739054e-02, -1.15877744e-02,  2.72059860e-03,  2.29104026e-03,\n",
      "       -9.97602474e-03, -9.76545364e-03,  1.99357723e-03, -3.25946696e-03,\n",
      "       -5.23123948e-04, -1.63604727e-03,  3.53714521e-03,  2.96805780e-02,\n",
      "        6.81886217e-04, -1.28719900e-02, -3.70872760e-04,  1.21853137e-02,\n",
      "        4.85449703e-03,  1.09190363e-02, -2.00749421e-03, -2.62271683e-03,\n",
      "        1.57380681e-02,  2.39197747e-04, -2.06456985e-02, -2.43829563e-03,\n",
      "       -3.33878724e-03,  4.43283841e-02,  3.46596731e-04, -1.54826313e-03,\n",
      "        6.88796444e-03,  1.74569841e-02,  2.80688126e-02,  2.52399743e-02,\n",
      "       -4.19081841e-03,  7.77495606e-03, -9.26667266e-03,  2.33864848e-04,\n",
      "        2.90688616e-03,  1.36379227e-02, -1.09948778e-04, -3.63345048e-03,\n",
      "        1.99716408e-02,  1.22022824e-02, -1.28422538e-02,  2.23179776e-02,\n",
      "        1.24540534e-02,  1.28550979e-03, -2.40548118e-03, -6.75575342e-03,\n",
      "       -1.22874286e-02, -7.00163329e-03, -2.28744908e-03, -1.35478651e-04,\n",
      "        6.79835863e-03,  8.27494729e-03,  1.56384706e-02, -6.05362980e-03,\n",
      "        1.09753094e-03,  3.66579578e-03, -1.92776043e-03,  1.30785359e-02,\n",
      "       -1.63714483e-03, -1.36339609e-02, -1.66218565e-03,  3.08781900e-02,\n",
      "        3.27514485e-02,  1.19507583e-02,  3.57903098e-03, -4.98172315e-03,\n",
      "        1.51888225e-02, -2.11248756e-04, -1.33969262e-02,  0.00000000e+00,\n",
      "        7.22540030e-03,  7.02545233e-03,  1.55843666e-03,  1.09359510e-02,\n",
      "       -6.32764446e-03, -5.70355542e-03, -5.51862828e-03, -8.76364205e-03,\n",
      "       -7.38237379e-03,  2.12407634e-02, -5.70349908e-03,  8.43972899e-03,\n",
      "        1.27869602e-02, -5.29799785e-04, -1.49689708e-03,  7.12978607e-03,\n",
      "        9.97270085e-03, -3.19647766e-03,  1.95367029e-03,  2.61215214e-02,\n",
      "        2.42057443e-02,  3.11903320e-02, -1.26208225e-02, -1.14809060e-02,\n",
      "       -1.35462796e-02, -4.78881644e-03,  4.01370344e-04, -1.36072384e-02,\n",
      "       -1.41785992e-03, -2.56886352e-02, -3.71606159e-03,  1.47856474e-02,\n",
      "       -1.08349381e-03, -2.22990289e-02,  4.54268930e-03, -1.05388314e-02,\n",
      "        2.18484327e-02,  1.64005198e-02,  1.76694721e-03, -4.32945741e-03,\n",
      "        6.34328322e-03, -3.35759460e-03, -1.14256199e-02,  1.49246817e-03,\n",
      "        0.00000000e+00, -1.66850328e-03, -1.44976459e-03,  2.42740922e-02,\n",
      "        1.67599022e-02, -3.52187641e-03,  3.26703973e-02, -7.65765039e-03,\n",
      "        7.39215780e-03,  5.85401477e-03,  9.75305494e-03, -2.05739215e-02,\n",
      "        2.54030759e-03,  6.19109953e-04,  1.21401437e-03, -1.64776202e-02,\n",
      "       -1.84858814e-02, -4.82080271e-03, -1.08980862e-02,  4.17542532e-02,\n",
      "       -1.46231186e-02, -1.26819368e-02,  2.49963794e-02,  1.84332784e-02,\n",
      "       -1.21981595e-02,  3.19569185e-02,  6.86419802e-03,  1.69537067e-02,\n",
      "       -3.29060084e-03,  6.78003344e-05, -1.73898775e-03,  1.23432288e-02,\n",
      "       -1.11196917e-02, -5.50628640e-03,  1.27771208e-02, -1.36032533e-02,\n",
      "        2.17404551e-04,  2.74847844e-03,  3.13359284e-04, -1.80826578e-02],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'padding': 'valid', 'data_format': 'channels_first', 'strides': (2, 2), 'pool_size': (2, 2), 'name': 'max_pooling2d_3', 'trainable': True} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'rate': 0.2, 'name': 'dropout_3', 'seed': None, 'trainable': True, 'noise_shape': None} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_size': (3, 3), 'name': 'conv2d_7', 'bias_regularizer': None, 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'filters': 512, 'use_bias': True, 'padding': 'same', 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[[[-1.78853534e-02,  1.84714347e-02, -2.80564055e-02, ...,\n",
      "           2.77510248e-02, -7.13862339e-03,  1.82964299e-02],\n",
      "         [-1.47807952e-02, -2.16459837e-02, -2.66617239e-02, ...,\n",
      "           1.05506293e-02, -2.99995206e-02, -5.73809678e-03],\n",
      "         [-9.55805089e-03,  2.51166523e-02, -2.16857940e-02, ...,\n",
      "           1.75652409e-03,  2.20875144e-02, -2.85846684e-02],\n",
      "         ...,\n",
      "         [-2.50161998e-02, -9.66587011e-03,  1.27904806e-02, ...,\n",
      "          -2.15245746e-02, -4.18024184e-03,  1.27224661e-02],\n",
      "         [-1.84739158e-02, -1.85572833e-03,  1.99504141e-02, ...,\n",
      "           2.46060770e-02, -1.66564609e-03, -2.56367233e-02],\n",
      "         [ 2.17413008e-02,  2.47329958e-02,  2.06031539e-02, ...,\n",
      "          -1.53959263e-02,  4.74984944e-03,  4.56395140e-03]],\n",
      "\n",
      "        [[-5.50964521e-03,  1.17447805e-02, -1.85027122e-02, ...,\n",
      "           7.66608492e-03,  5.59023162e-03,  3.05408612e-03],\n",
      "         [-2.34752847e-03,  4.96713910e-03,  2.10060515e-02, ...,\n",
      "           2.07896009e-02, -1.88548304e-02,  2.45687775e-02],\n",
      "         [ 1.17812715e-02,  2.71887947e-02,  2.44781282e-02, ...,\n",
      "           1.99891534e-03, -1.20136579e-02,  1.54175153e-02],\n",
      "         ...,\n",
      "         [ 1.84118543e-02,  2.88699269e-02,  1.74485855e-02, ...,\n",
      "           2.17144582e-02, -2.87709534e-02, -5.71860699e-03],\n",
      "         [ 2.86784302e-02,  5.76948281e-03, -7.10861059e-03, ...,\n",
      "           2.61965133e-02,  2.87900474e-02, -7.34814536e-03],\n",
      "         [-2.21387167e-02, -1.95273068e-02, -2.64485250e-03, ...,\n",
      "           2.04672441e-02,  3.88551620e-03, -1.31318662e-02]],\n",
      "\n",
      "        [[-2.88858842e-02,  2.34618839e-02,  2.58472096e-02, ...,\n",
      "          -2.49969270e-02,  1.60713457e-02, -2.71927249e-02],\n",
      "         [-2.52629723e-02, -2.47053038e-02, -9.11999494e-03, ...,\n",
      "           4.18274477e-03,  1.63580738e-02,  2.35477034e-02],\n",
      "         [ 1.32691674e-02,  1.44961793e-02, -1.37414439e-02, ...,\n",
      "           1.43428957e-02, -3.68714076e-03, -2.23918166e-02],\n",
      "         ...,\n",
      "         [-2.61172242e-02, -2.50782985e-02,  1.74530502e-02, ...,\n",
      "          -1.34645244e-02,  1.18337767e-02, -3.37730604e-03],\n",
      "         [ 7.38397054e-03,  1.68807395e-02,  5.45892213e-03, ...,\n",
      "           2.89278850e-02, -2.86555471e-04,  2.50352323e-02],\n",
      "         [-3.98600195e-03, -2.98645236e-02, -2.34764386e-02, ...,\n",
      "           2.30412651e-02,  2.27434244e-02, -2.82364823e-02]]],\n",
      "\n",
      "\n",
      "       [[[-2.86526065e-02, -2.93749403e-02, -1.70901120e-02, ...,\n",
      "           2.30104886e-02, -2.79210601e-02, -1.37008848e-02],\n",
      "         [-2.67200079e-02,  2.49963328e-02, -2.98012746e-03, ...,\n",
      "           1.83282848e-02, -8.82641133e-03, -9.70003475e-03],\n",
      "         [ 2.37656906e-02,  1.56151270e-02,  1.70887075e-03, ...,\n",
      "           3.96006973e-03, -2.09694542e-02, -1.02706794e-02],\n",
      "         ...,\n",
      "         [ 2.83030774e-02, -1.35284606e-02,  2.85853893e-02, ...,\n",
      "          -7.07236445e-03,  3.62919853e-03,  7.02019362e-03],\n",
      "         [-5.44055970e-03,  3.99301248e-03, -1.51105253e-02, ...,\n",
      "           2.17451137e-02,  5.24401478e-03,  2.10721549e-02],\n",
      "         [ 1.25627080e-02, -1.03478669e-03, -1.08748768e-02, ...,\n",
      "           7.57587142e-03,  8.98436829e-03, -2.54189838e-02]],\n",
      "\n",
      "        [[ 7.44872540e-03,  1.68786626e-02,  5.49844978e-03, ...,\n",
      "           1.40274651e-02, -1.30193690e-02,  4.01057582e-03],\n",
      "         [-1.06775276e-02, -2.16876958e-02,  1.38124442e-02, ...,\n",
      "          -2.26379419e-03, -6.86908548e-04, -9.26188752e-03],\n",
      "         [ 7.73190800e-03, -1.19119892e-02,  2.57223099e-02, ...,\n",
      "           1.91220660e-02,  2.78188661e-02, -1.93778165e-02],\n",
      "         ...,\n",
      "         [ 2.22224575e-02, -1.19508142e-02, -4.07999009e-03, ...,\n",
      "          -1.45742930e-02, -2.95529924e-02,  9.70306061e-03],\n",
      "         [-2.03122497e-02, -2.57014483e-02,  8.87161307e-03, ...,\n",
      "           1.95682831e-02, -2.60858843e-03, -2.81712301e-02],\n",
      "         [-1.44173885e-02,  1.42351026e-02,  9.63803206e-04, ...,\n",
      "           1.14020612e-02, -2.07416266e-02, -2.06349138e-03]],\n",
      "\n",
      "        [[ 2.11835448e-02, -6.87452592e-03, -8.76444275e-04, ...,\n",
      "          -7.52499653e-03, -1.63632724e-02, -4.56834491e-03],\n",
      "         [ 6.35855319e-03, -1.95770375e-02, -2.09472179e-02, ...,\n",
      "           1.79267358e-02, -2.21172553e-02,  8.96312110e-03],\n",
      "         [ 8.97564925e-03,  1.26544572e-02,  1.74140949e-02, ...,\n",
      "          -1.97091028e-02,  1.53346900e-02, -1.44968312e-02],\n",
      "         ...,\n",
      "         [-9.21901502e-03, -9.81130637e-03,  2.22936589e-02, ...,\n",
      "          -7.55022885e-03,  2.44991705e-02,  1.76035538e-02],\n",
      "         [ 8.05186294e-03,  2.29000188e-02, -2.04201303e-02, ...,\n",
      "          -6.49323082e-03, -1.30087622e-02,  1.23281628e-02],\n",
      "         [-1.77136641e-02,  1.07410075e-02, -2.15570126e-02, ...,\n",
      "           1.28002279e-02, -1.19786486e-02, -1.24484738e-02]]],\n",
      "\n",
      "\n",
      "       [[[-2.65501291e-02,  3.83135607e-03,  1.39976330e-02, ...,\n",
      "           2.57942937e-02,  1.56768598e-02,  1.69164613e-02],\n",
      "         [-2.23566368e-02, -1.98480226e-02,  1.19184330e-02, ...,\n",
      "          -8.45155399e-03,  2.06963159e-02, -2.17095949e-02],\n",
      "         [-6.88183354e-03, -8.42960365e-03, -1.82891022e-02, ...,\n",
      "           1.66630503e-02,  2.40503848e-02, -2.64753364e-02],\n",
      "         ...,\n",
      "         [-1.34240240e-02, -2.74100825e-02, -1.06617976e-02, ...,\n",
      "           2.07582451e-02,  1.90714300e-02,  2.16834564e-02],\n",
      "         [-2.34069843e-02,  2.41499282e-02,  7.46333506e-03, ...,\n",
      "           2.29826886e-02,  2.40803119e-02, -2.50786487e-02],\n",
      "         [ 1.51257031e-03, -1.44952498e-02, -1.84541699e-02, ...,\n",
      "           1.46726249e-02,  1.54476345e-03,  2.34023179e-03]],\n",
      "\n",
      "        [[ 2.52699312e-02, -5.86613896e-05,  1.13924649e-02, ...,\n",
      "           2.15729903e-02,  2.62163542e-02, -1.04806665e-02],\n",
      "         [-1.10085830e-02,  1.56687740e-02, -8.40689987e-03, ...,\n",
      "           2.18002824e-03,  5.07725496e-03,  3.78583977e-03],\n",
      "         [-2.72621009e-02,  1.03116445e-02,  1.45269549e-02, ...,\n",
      "           9.27779358e-03, -2.01270338e-02,  1.42129865e-02],\n",
      "         ...,\n",
      "         [-2.75994055e-02,  6.22710679e-03, -4.43470944e-03, ...,\n",
      "           9.99004813e-04,  9.63674672e-03,  2.69613694e-02],\n",
      "         [ 2.41675228e-02,  2.85188183e-02, -2.42519137e-02, ...,\n",
      "           1.36385355e-02, -8.77498183e-03,  9.74144728e-04],\n",
      "         [ 2.69498918e-02, -3.04699391e-02, -1.41656585e-02, ...,\n",
      "          -1.76890325e-02, -2.17855349e-02, -1.10205915e-02]],\n",
      "\n",
      "        [[ 2.00746022e-02, -2.00621821e-02, -8.30714498e-03, ...,\n",
      "           2.38265824e-02,  6.09841989e-03, -2.08112775e-04],\n",
      "         [ 1.58707779e-02, -2.58526821e-02, -2.30494346e-02, ...,\n",
      "          -2.28048279e-03, -1.11444611e-02, -1.58227831e-02],\n",
      "         [-1.51289310e-02,  2.51164547e-05,  2.17842124e-02, ...,\n",
      "          -7.00919889e-04, -3.12251505e-03,  1.87208550e-03],\n",
      "         ...,\n",
      "         [ 1.15102688e-02,  5.53845800e-03, -2.57668849e-02, ...,\n",
      "           2.27363836e-02,  2.60080658e-02,  7.66956538e-04],\n",
      "         [-2.79407520e-02, -2.20887805e-03,  1.57029473e-03, ...,\n",
      "          -1.33705316e-02, -2.68099532e-02, -1.39537528e-02],\n",
      "         [-9.45868809e-03, -1.88411714e-03, -3.73190502e-03, ...,\n",
      "          -2.79307477e-02,  2.23029265e-03,  9.49031208e-03]]]],\n",
      "      dtype=float32), array([ 6.98323548e-03,  2.57695243e-02,  3.60670337e-03, -1.18340058e-02,\n",
      "        6.60752319e-03, -4.79733804e-03, -1.45099908e-02, -1.88722659e-03,\n",
      "        1.68414786e-04,  2.51360126e-02,  4.26930841e-03, -3.93398618e-03,\n",
      "        2.09151432e-02, -1.49987300e-03,  8.00154358e-03,  9.72350501e-03,\n",
      "        1.54024025e-03,  2.79010390e-03, -3.17687578e-02,  1.84013613e-03,\n",
      "       -4.32938337e-03,  3.96391610e-03,  8.26954190e-03, -2.56884880e-02,\n",
      "        1.24758463e-02,  2.79666651e-02,  5.10114944e-04,  2.46704333e-02,\n",
      "       -3.34826158e-03, -3.99112329e-03, -2.19208351e-03,  2.84920260e-03,\n",
      "       -9.57448874e-03,  2.43497230e-02,  3.20891440e-02, -6.83734761e-05,\n",
      "       -4.87602549e-03, -1.39116235e-02,  5.87570015e-04,  1.45591199e-02,\n",
      "       -1.79353345e-03,  1.82834622e-02, -2.12594474e-04,  1.66506656e-02,\n",
      "       -7.37508852e-03,  4.39332891e-03, -4.91920998e-03,  3.07012000e-04,\n",
      "        4.70520835e-03,  1.02881119e-02,  1.18099097e-02,  4.09855135e-03,\n",
      "        9.68954642e-04,  3.08827264e-03,  3.37721501e-03, -5.21243270e-03,\n",
      "       -5.42129856e-03, -4.73254267e-03, -4.43658186e-03, -1.44192306e-02,\n",
      "        3.09544499e-03, -9.78937373e-04,  5.67092421e-03,  3.54828808e-04,\n",
      "       -9.42696817e-03, -4.25228011e-03, -3.60188307e-03, -4.59923502e-03,\n",
      "       -9.88612301e-04, -1.00687752e-03, -3.92353220e-04, -3.15595255e-03,\n",
      "        4.27377671e-02,  3.43147106e-02,  4.99669136e-03,  1.33760842e-02,\n",
      "       -3.57289170e-03,  1.18274763e-02, -2.86234007e-03,  7.17447326e-03,\n",
      "        1.07658748e-03,  1.02721946e-02,  1.56579185e-02, -3.13498313e-04,\n",
      "        1.46288308e-03, -1.56278850e-03,  7.19975587e-03,  7.23349722e-03,\n",
      "        5.30320511e-04, -1.46262941e-03, -2.69768178e-03,  3.47530411e-04,\n",
      "       -3.13485810e-03, -9.50491056e-03,  3.05996655e-04,  2.19336171e-02,\n",
      "        3.39949690e-02,  1.37198269e-02, -1.55607981e-04,  8.65642633e-03,\n",
      "        1.12805155e-03,  1.08531611e-02, -2.98116566e-03,  6.38006895e-04,\n",
      "       -1.41451079e-02,  1.24756796e-02,  5.65671781e-03, -5.32482564e-03,\n",
      "        8.51551071e-03,  2.20629200e-02, -6.66910782e-03, -3.61539162e-04,\n",
      "        2.13113660e-03, -1.20357359e-02,  4.44761151e-03, -9.45975911e-03,\n",
      "       -1.10297427e-02, -1.05604100e-04,  1.94659065e-02,  1.52240451e-02,\n",
      "       -2.17419863e-03, -7.73688685e-03,  3.71054746e-03, -9.90847964e-03,\n",
      "        3.05021610e-02, -3.43807833e-03,  3.54822204e-02,  1.61176641e-02,\n",
      "        2.71737692e-03,  1.01258773e-02, -9.97063052e-03,  1.58071127e-02,\n",
      "       -1.72051545e-02, -6.59977691e-03, -1.32134918e-03, -1.47275850e-02,\n",
      "       -3.31297540e-03,  5.86448126e-02, -1.16282806e-03,  2.27024052e-02,\n",
      "       -1.12172524e-02, -2.18073861e-03, -3.54538381e-04,  1.59033574e-03,\n",
      "        2.32454133e-03, -2.44500190e-02, -4.15953109e-04,  4.28566569e-03,\n",
      "        4.06822376e-03,  2.99417749e-02,  4.06953460e-03, -3.39056831e-03,\n",
      "        4.30452172e-03,  4.24410701e-02,  3.81154194e-02,  5.48155606e-03,\n",
      "        1.40771794e-04,  8.51636752e-04, -4.28434415e-03,  1.49851048e-03,\n",
      "       -1.13144726e-03, -8.10190197e-03,  1.64312881e-03,  1.62015355e-03,\n",
      "        2.25560851e-02, -9.88437887e-03, -2.26437510e-03,  2.68060551e-03,\n",
      "       -8.38012248e-03,  7.17236428e-04,  1.95326265e-02, -6.25018554e-04,\n",
      "        2.62338639e-04, -8.94952158e-04,  3.68278101e-03,  5.89521416e-03,\n",
      "       -8.64092540e-03, -1.03843259e-02, -1.31809921e-03,  8.31034407e-03,\n",
      "        5.16869090e-02, -1.11344792e-02, -6.29764574e-04, -1.56290852e-03,\n",
      "        5.88104129e-03, -1.11909129e-03, -8.42919480e-03,  2.06453521e-02,\n",
      "       -3.91769968e-03,  2.68622045e-03,  6.94088405e-03, -4.78019845e-03,\n",
      "       -5.89476852e-03,  9.73398890e-03, -8.85358080e-03,  1.59920305e-02,\n",
      "       -7.92855956e-03,  8.86440184e-03, -1.50377993e-02, -1.45168300e-03,\n",
      "        4.80881408e-02,  4.78076795e-03, -1.23604108e-03,  1.40291487e-03,\n",
      "        3.02656990e-04,  6.12265989e-03, -5.29313926e-03,  1.45988353e-03,\n",
      "       -4.13760729e-03,  3.49715771e-03,  8.45735986e-03,  2.20049452e-02,\n",
      "       -9.80631169e-03, -4.83912445e-04,  3.47708818e-03,  8.39464180e-03,\n",
      "       -1.54492846e-02, -1.59329847e-02, -1.02861123e-02, -7.28979055e-03,\n",
      "        1.19265830e-02,  2.18062848e-03,  4.19324869e-03, -2.83089709e-02,\n",
      "       -2.74529075e-03,  1.94948893e-02, -1.07191876e-02, -4.71567386e-04,\n",
      "        6.71607535e-03, -7.77843874e-03, -1.48243457e-03, -1.03084685e-03,\n",
      "       -9.48797446e-03,  1.08335521e-02,  7.54953688e-03, -1.03281299e-03,\n",
      "        4.60859425e-02, -2.96271103e-03, -3.45275825e-04, -4.81366878e-03,\n",
      "        5.89432232e-02,  1.41920857e-02,  3.14054516e-04,  9.93148307e-04,\n",
      "       -4.53493698e-03, -8.37752596e-03,  9.53950919e-03, -4.13548294e-03,\n",
      "       -1.10314526e-02, -6.18513953e-03,  1.29165000e-03, -7.22582918e-03,\n",
      "        6.62979158e-03, -5.32956328e-03,  2.14843154e-02,  3.76904290e-03,\n",
      "        1.60124637e-02, -3.72332730e-03,  8.77054466e-04,  4.47926781e-04,\n",
      "        1.21007487e-02,  1.31098805e-02,  2.61275703e-03,  2.84174625e-02,\n",
      "        9.39499005e-04,  5.30620012e-03,  1.28650441e-04,  1.08287930e-02,\n",
      "       -2.05672029e-02, -1.30938282e-02, -5.59223164e-03,  2.17077620e-02,\n",
      "        4.57611168e-03,  5.02397027e-03, -1.59175079e-02,  9.59959230e-04,\n",
      "        8.91312957e-03,  6.22503646e-03, -3.93891009e-03,  1.21089313e-02,\n",
      "       -4.13986994e-03, -2.99511082e-03, -4.60076937e-03, -1.84846129e-02,\n",
      "       -1.80688151e-03, -1.93471881e-03,  5.13129719e-02, -8.02788883e-03,\n",
      "        5.20452373e-02,  1.94175150e-02,  2.78984439e-02,  5.46039315e-03,\n",
      "        1.58147514e-02,  1.44012726e-03,  1.12311682e-04, -2.17912383e-02,\n",
      "       -2.64830375e-03,  5.59080578e-03,  3.21080070e-03, -1.04443142e-02,\n",
      "        2.79649794e-02,  4.00432339e-03,  8.59882589e-03,  1.14094943e-03,\n",
      "       -1.41950930e-02,  2.39187628e-02, -2.75153172e-04,  1.05018048e-02,\n",
      "       -8.50472040e-03,  1.48794148e-03, -1.02278246e-02,  1.85562279e-02,\n",
      "       -1.14053795e-02,  1.38445096e-02,  1.69643126e-02, -9.12543107e-03,\n",
      "       -3.45037282e-02,  2.37321593e-02,  2.00051372e-03, -1.19361402e-02,\n",
      "        5.39365783e-03,  3.57903726e-02,  5.05423546e-03, -1.16895381e-02,\n",
      "        2.44709011e-02, -2.07733624e-02, -1.95245713e-03,  2.03294177e-02,\n",
      "       -1.44313844e-02,  6.07719237e-04, -8.79455823e-03,  2.07001958e-02,\n",
      "       -1.25561068e-02,  3.78620205e-03, -3.12148943e-03, -1.17752359e-04,\n",
      "        5.08881314e-03, -1.86012946e-02,  1.09422747e-02,  7.74655724e-03,\n",
      "       -8.81846529e-03, -7.75626069e-03,  8.25639348e-04,  4.85491328e-04,\n",
      "       -1.95625331e-03,  6.78422209e-03, -1.98381007e-04, -1.36553252e-03,\n",
      "       -2.52911169e-03,  8.60051438e-03,  4.54297923e-02,  3.84251364e-02,\n",
      "        2.16928124e-02, -9.15980339e-03, -1.45853916e-03,  3.87160713e-03,\n",
      "       -2.64126132e-03, -9.01514868e-05, -1.11876344e-02,  1.29026053e-02,\n",
      "       -8.98581930e-04,  7.26659084e-03,  6.42558411e-02,  9.48889274e-03,\n",
      "        7.18266796e-03,  2.98180641e-03, -9.39668529e-03,  1.67555008e-02,\n",
      "       -3.80490976e-03, -8.53877806e-04,  6.38811104e-03,  3.13196518e-03,\n",
      "        5.14688622e-03, -1.33280747e-03,  4.39893641e-03,  1.52169857e-02,\n",
      "        6.49279496e-03, -2.65121507e-03,  3.53185460e-02, -9.06945858e-03,\n",
      "       -3.57447378e-03,  1.74693540e-02, -1.95442419e-03, -2.12371722e-03,\n",
      "       -9.85482149e-03,  1.26091680e-02, -1.23011665e-02,  1.00686233e-02,\n",
      "       -8.84872302e-03, -3.72918369e-03,  6.79042377e-03,  3.56319174e-02,\n",
      "       -1.80464461e-02, -3.59785906e-03, -1.64676376e-03, -4.59582778e-03,\n",
      "       -2.20958522e-04, -5.29473275e-03,  2.05200687e-02,  2.14579329e-03,\n",
      "       -1.09129883e-02,  2.74810605e-02, -3.83586925e-03,  3.84597341e-04,\n",
      "       -1.04172546e-02,  1.22050820e-02, -1.32229761e-03, -3.86034860e-03,\n",
      "        3.58681253e-04, -3.86146735e-03,  8.06027791e-04, -2.01348434e-04,\n",
      "       -1.03420774e-02,  1.22897117e-03,  5.23715606e-03,  4.13341969e-02,\n",
      "        6.70337467e-04, -1.15005685e-04, -1.57222450e-02, -1.93805096e-03,\n",
      "        1.10647269e-02,  6.74040639e-04, -4.24441602e-03, -1.11017926e-02,\n",
      "        1.06749320e-02, -1.98044977e-03, -8.40779487e-03,  5.06258570e-02,\n",
      "        6.73066229e-02, -1.04113491e-02,  3.12863849e-02, -9.19796526e-03,\n",
      "       -1.15861213e-02,  7.75482622e-04,  1.72790438e-02,  1.16986834e-04,\n",
      "        1.40449684e-03, -3.83321638e-03, -6.49360649e-04, -3.32538295e-03,\n",
      "        8.50993558e-04,  8.95309914e-03,  4.25694371e-03,  1.78305656e-02,\n",
      "        1.60065051e-02,  4.54853941e-03,  5.53684123e-03,  1.14094783e-02,\n",
      "       -1.61947345e-03, -4.74292971e-03, -4.45488468e-03,  5.63728437e-02,\n",
      "       -2.71650637e-03, -9.78046912e-04, -9.50947404e-03,  6.40866440e-03,\n",
      "        1.93943596e-03, -7.52484938e-03,  7.99193606e-03, -6.40423782e-03,\n",
      "       -1.38563449e-02, -2.12223316e-03, -1.32805726e-03, -1.75791252e-02,\n",
      "       -3.61288595e-03, -7.23388279e-03,  3.03954352e-04, -2.58650817e-02,\n",
      "       -2.52668951e-02, -8.57132021e-03, -3.37727834e-03, -1.67685393e-02,\n",
      "       -8.25102814e-03,  1.26541043e-02,  1.02222254e-02, -2.90788501e-03,\n",
      "        4.99607017e-03,  1.88331362e-02,  6.11954508e-03,  1.60391480e-02,\n",
      "        3.52397165e-03, -2.15735137e-02, -8.91685486e-03, -2.20450433e-03,\n",
      "        8.27741530e-03,  5.71885030e-04,  1.55072089e-03,  9.35571268e-03,\n",
      "        9.45332285e-04,  5.45714162e-02, -4.35359031e-03,  1.44510772e-02,\n",
      "        5.34707159e-02, -3.24559119e-03,  8.18177778e-03,  2.96399444e-02,\n",
      "       -3.35312709e-02, -3.67509900e-03,  9.35134385e-03, -1.79378241e-02,\n",
      "        9.51825990e-04, -7.52348162e-04, -7.05580739e-03,  1.33036822e-03,\n",
      "       -1.80762298e-02, -1.28850201e-03,  2.64058588e-03,  1.71730332e-02,\n",
      "        2.40477156e-02, -4.51436499e-03, -8.50970484e-03,  6.82231411e-03],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "g==  {'data_format': 'channels_first', 'strides': (1, 1), 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'dilation_rate': (1, 1), 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'kernel_size': (3, 3), 'name': 'conv2d_8', 'bias_regularizer': None, 'bias_constraint': None, 'trainable': True, 'activation': 'relu', 'activity_regularizer': None, 'filters': 512, 'use_bias': True, 'padding': 'valid', 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[[[ 8.31422210e-03,  3.14529333e-03, -1.39101604e-02, ...,\n",
      "          -9.76064429e-03,  1.93997342e-02, -1.95101905e-03],\n",
      "         [ 9.25896503e-03,  1.61468461e-02,  5.68474224e-03, ...,\n",
      "           1.58274826e-02, -1.45675540e-02, -1.73643057e-03],\n",
      "         [ 6.26318724e-05,  1.67501706e-03,  3.66939977e-03, ...,\n",
      "          -3.19525339e-02, -7.63526093e-03, -2.48654224e-02],\n",
      "         ...,\n",
      "         [ 1.64923239e-02,  8.50984268e-03, -1.41956843e-03, ...,\n",
      "          -3.31903662e-04,  1.28697585e-02, -1.38411596e-02],\n",
      "         [ 2.25150101e-02, -1.37651209e-02,  2.34302189e-02, ...,\n",
      "          -9.33973957e-03,  4.12291242e-03, -1.85219049e-02],\n",
      "         [ 1.85786970e-02,  9.69809573e-03,  2.06450503e-02, ...,\n",
      "           2.43923441e-03, -2.71492675e-02,  2.09036935e-02]],\n",
      "\n",
      "        [[-2.26299819e-02, -1.24987531e-02,  2.45750719e-03, ...,\n",
      "           1.62956249e-02,  6.37269393e-03,  1.84377376e-02],\n",
      "         [ 1.78914182e-02,  1.12736868e-02, -6.62511261e-03, ...,\n",
      "           1.17418347e-02, -8.71486124e-03, -1.30169597e-02],\n",
      "         [-8.86724517e-03, -1.83266252e-02, -1.40236197e-02, ...,\n",
      "          -9.53781605e-03,  2.22673044e-02, -1.93594974e-02],\n",
      "         ...,\n",
      "         [-4.06649575e-04,  9.50567145e-03, -2.15322282e-02, ...,\n",
      "           7.44136469e-03,  1.99828949e-02,  7.95390643e-03],\n",
      "         [-1.37020648e-02,  1.81415174e-02, -1.82507560e-02, ...,\n",
      "          -1.44466730e-02, -3.27784568e-03,  9.50933993e-03],\n",
      "         [-2.32314281e-02,  2.49049645e-02,  5.51764295e-03, ...,\n",
      "          -2.69433036e-02, -1.91509631e-02,  6.75683422e-03]],\n",
      "\n",
      "        [[ 4.67199320e-03,  4.49932227e-03,  4.77519818e-03, ...,\n",
      "           2.16402505e-02, -2.22394541e-02, -2.83489213e-03],\n",
      "         [-2.50546932e-02, -1.80452410e-02,  1.45612517e-02, ...,\n",
      "           1.01711266e-02, -1.02835326e-02,  9.75603587e-04],\n",
      "         [-4.32740012e-03,  5.66588528e-03,  6.37320057e-03, ...,\n",
      "          -8.21914524e-03,  2.37336922e-02, -3.65139171e-03],\n",
      "         ...,\n",
      "         [-1.56501234e-02, -7.60783115e-03,  2.20430177e-02, ...,\n",
      "          -1.23729976e-02,  1.80498827e-02, -1.61968507e-02],\n",
      "         [-1.04246698e-02, -1.83899645e-02, -2.44365502e-02, ...,\n",
      "          -2.81496393e-03, -1.75119489e-02,  1.80394668e-02],\n",
      "         [-1.29484031e-02,  1.03796860e-02, -4.28503379e-03, ...,\n",
      "          -2.36819442e-02,  1.12699633e-02,  1.16203623e-02]]],\n",
      "\n",
      "\n",
      "       [[[-1.08431373e-02, -2.43476983e-02, -2.42127441e-02, ...,\n",
      "          -1.07223131e-02,  6.50139584e-04,  4.00656881e-03],\n",
      "         [-9.13112611e-03, -5.61643101e-04, -1.79719124e-02, ...,\n",
      "          -2.37745233e-02, -2.41371915e-02,  8.36518779e-03],\n",
      "         [-2.51650140e-02,  7.84110278e-03,  1.52249373e-02, ...,\n",
      "           7.82164373e-03, -7.61569897e-03, -1.38137918e-02],\n",
      "         ...,\n",
      "         [-1.28088873e-02,  8.29278305e-03,  2.09775511e-02, ...,\n",
      "           1.28765991e-02, -3.62333213e-03, -2.26972606e-02],\n",
      "         [ 2.34225895e-02,  2.34011635e-02,  9.92217753e-03, ...,\n",
      "          -3.26891206e-02,  1.98974609e-02, -1.55175636e-02],\n",
      "         [ 6.98309857e-03, -1.56389270e-02, -2.46615242e-02, ...,\n",
      "           7.25900847e-03,  8.84335767e-03, -1.52813019e-02]],\n",
      "\n",
      "        [[ 1.03282295e-02, -9.05731414e-03,  1.80927839e-03, ...,\n",
      "           2.02861857e-02,  1.83424763e-02, -8.06749705e-03],\n",
      "         [-2.61096377e-02, -1.58771612e-02,  1.88200492e-02, ...,\n",
      "           1.02042174e-02,  9.35314316e-03, -1.62696801e-02],\n",
      "         [ 1.53017957e-02, -2.03422178e-02,  1.87703930e-02, ...,\n",
      "           1.96104823e-03,  1.37107056e-02,  1.18961306e-02],\n",
      "         ...,\n",
      "         [ 4.17341304e-04,  2.82712700e-03, -2.60608941e-02, ...,\n",
      "          -1.94886532e-02,  2.34485208e-03, -1.88523047e-02],\n",
      "         [-2.70543946e-03, -1.07174311e-02, -1.78925283e-02, ...,\n",
      "           7.23110279e-03, -8.13882519e-03,  1.43105099e-02],\n",
      "         [-2.48284228e-02, -1.41678741e-02, -1.34078255e-02, ...,\n",
      "          -1.41988066e-03,  7.62108853e-03,  6.87462231e-03]],\n",
      "\n",
      "        [[-2.00174972e-02,  1.46127050e-03, -2.40299990e-03, ...,\n",
      "          -2.03276873e-02, -2.00748034e-02, -1.15194907e-02],\n",
      "         [-1.67880841e-02, -2.37470213e-02,  2.22266326e-03, ...,\n",
      "           1.56153338e-02, -8.80394038e-03,  2.18090322e-02],\n",
      "         [ 2.00818647e-02, -1.50139565e-02,  1.34744188e-02, ...,\n",
      "           7.72220036e-03,  4.36451845e-03,  2.07631048e-02],\n",
      "         ...,\n",
      "         [-1.75525490e-02, -1.03408098e-02,  4.36113728e-03, ...,\n",
      "           1.21709462e-02,  2.00360585e-02, -4.30724723e-03],\n",
      "         [ 1.14651462e-02, -9.55034606e-03, -2.37454697e-02, ...,\n",
      "          -1.86044537e-02, -2.33769324e-02,  1.59551501e-02],\n",
      "         [ 1.42729925e-02, -2.11744923e-02, -3.86811630e-03, ...,\n",
      "          -1.92492194e-02, -1.87240019e-02, -1.29700331e-02]]],\n",
      "\n",
      "\n",
      "       [[[ 1.01998104e-02, -1.09828543e-02,  1.94670931e-02, ...,\n",
      "          -2.47099553e-03, -9.73189529e-03, -2.25575473e-02],\n",
      "         [ 8.78904015e-03,  1.66587140e-02, -6.31858408e-03, ...,\n",
      "          -1.37467664e-02, -1.61941163e-02,  2.53249146e-03],\n",
      "         [-1.51122669e-02,  1.47526516e-02,  2.27855779e-02, ...,\n",
      "          -1.59480404e-02, -2.94879195e-03, -9.27984156e-03],\n",
      "         ...,\n",
      "         [-1.50854737e-02,  1.50746887e-03, -1.05835265e-02, ...,\n",
      "          -1.64253507e-02,  3.96383042e-03,  1.96517669e-02],\n",
      "         [ 8.20539519e-03, -2.82813242e-04,  3.46753257e-03, ...,\n",
      "          -2.15055589e-02,  1.47981299e-02, -2.40912661e-02],\n",
      "         [ 1.38967223e-02,  1.78922229e-02, -1.19681051e-02, ...,\n",
      "          -1.86618641e-02, -1.29852910e-02,  2.45869756e-02]],\n",
      "\n",
      "        [[-9.62680764e-03, -2.33025532e-02, -1.69087760e-02, ...,\n",
      "          -3.40662268e-03,  1.93243455e-02, -2.49738414e-02],\n",
      "         [ 1.32494876e-02,  9.13554057e-03,  1.90539956e-02, ...,\n",
      "           1.65040698e-02,  1.29501354e-02,  7.12872576e-03],\n",
      "         [-4.53525223e-03, -2.27568559e-02, -2.17003822e-02, ...,\n",
      "           9.86840483e-03, -4.11802158e-03,  5.93369547e-03],\n",
      "         ...,\n",
      "         [ 2.04425659e-02, -1.68145951e-02, -3.16652330e-03, ...,\n",
      "           9.69486963e-03,  2.21426301e-02,  1.79755520e-02],\n",
      "         [ 1.87598560e-02, -8.96321516e-03,  1.74349919e-02, ...,\n",
      "          -2.29452979e-02, -2.50492264e-02,  2.25733556e-02],\n",
      "         [-1.50076561e-02,  9.57849994e-03, -1.65719520e-02, ...,\n",
      "           1.65391192e-02, -2.49516945e-02,  1.65429506e-02]],\n",
      "\n",
      "        [[-2.45880131e-02,  8.04441329e-03,  1.18875159e-02, ...,\n",
      "          -1.91456992e-02, -9.31525510e-03, -2.13991459e-02],\n",
      "         [ 8.83891527e-03, -6.26614725e-04,  2.20955908e-02, ...,\n",
      "           1.95165928e-02, -2.23769993e-02, -2.02003624e-02],\n",
      "         [ 1.86358881e-03, -2.38444433e-02,  1.01248054e-02, ...,\n",
      "           1.59114189e-02, -6.31344272e-03, -1.55655257e-02],\n",
      "         ...,\n",
      "         [ 1.39982216e-02, -2.02478617e-02,  1.85186695e-02, ...,\n",
      "           1.00485086e-02, -1.39402915e-02, -1.59629621e-02],\n",
      "         [ 2.00584419e-02, -2.18468644e-02,  5.82373468e-03, ...,\n",
      "           1.15620419e-02, -1.85532328e-02, -2.83295033e-03],\n",
      "         [ 1.16184596e-02,  4.42609703e-03,  8.67715850e-03, ...,\n",
      "          -1.20010944e-02,  1.26425205e-02,  5.51794423e-03]]]],\n",
      "      dtype=float32), array([ 1.45490747e-03, -7.91889057e-03,  3.76721248e-02, -1.39306975e-03,\n",
      "       -1.73682608e-02,  2.72528897e-03, -1.01728272e-02, -2.86871474e-02,\n",
      "        2.63169617e-03,  2.48831455e-02,  1.10583836e-02, -1.32854306e-03,\n",
      "       -4.31783078e-03,  1.42739026e-03, -6.54036971e-03, -1.01572303e-02,\n",
      "       -6.13786932e-03,  1.94541272e-02,  2.12667566e-02, -2.49419007e-02,\n",
      "        2.34402232e-02, -3.94696184e-03, -2.79102183e-04, -1.89643987e-02,\n",
      "        4.88936156e-03,  2.88076103e-02, -1.76762827e-02,  6.73054194e-04,\n",
      "       -5.42387739e-03,  1.88897122e-02,  8.76486581e-03,  5.31822145e-02,\n",
      "       -3.25955055e-03, -1.61123853e-02, -8.60194582e-03,  3.25938105e-03,\n",
      "        1.94692258e-02,  2.21991888e-03, -3.11347004e-03,  2.76042093e-02,\n",
      "       -2.00724900e-02, -1.25328917e-03, -7.24240625e-03,  4.13283780e-02,\n",
      "       -4.36196523e-03, -1.35751795e-02,  1.42852012e-02,  1.09715136e-02,\n",
      "        2.16209167e-03,  1.04617057e-02, -2.24381289e-03,  7.09853275e-03,\n",
      "        5.95813766e-02,  1.16371457e-02,  1.49199463e-04,  8.10972997e-04,\n",
      "        4.14988399e-03,  1.42734917e-02, -3.87984538e-03,  1.78287011e-02,\n",
      "       -2.70149740e-03,  1.58052072e-02, -1.95993762e-03,  2.38667335e-03,\n",
      "       -6.98482501e-04, -3.17493541e-05,  3.76703683e-03,  9.92643367e-03,\n",
      "       -1.02155395e-02,  4.32467833e-02, -5.55666815e-03,  1.38959717e-02,\n",
      "        1.72956358e-03,  9.60331783e-03,  6.83521619e-03, -4.70627844e-03,\n",
      "       -9.03082266e-03,  5.98690379e-03,  2.85870135e-02, -1.31994486e-02,\n",
      "       -9.90918092e-03,  1.08894368e-03,  4.78532398e-03,  6.52263872e-03,\n",
      "        3.35705318e-02, -2.56473571e-03, -2.41737091e-03, -4.62017860e-03,\n",
      "        1.28759863e-02, -3.68881933e-02, -5.37028676e-03, -1.18328456e-03,\n",
      "        7.45959580e-03, -1.51479419e-03, -1.50167709e-02,  4.30315435e-02,\n",
      "       -2.30177189e-03, -1.28319431e-02,  3.14438045e-02, -2.82744202e-03,\n",
      "        8.26006941e-03, -1.72321238e-02,  1.17150566e-03, -1.03297383e-02,\n",
      "        5.20751551e-02,  9.74959228e-03, -2.15858730e-04, -2.63397349e-03,\n",
      "        4.74204728e-03,  1.73660088e-02,  6.30655233e-03,  2.53941938e-02,\n",
      "        2.40887217e-02,  8.89919885e-03,  4.10785824e-02, -2.58713914e-03,\n",
      "       -1.10125123e-03, -1.27403755e-02, -2.44737472e-02, -2.98699015e-03,\n",
      "        3.40452716e-02, -5.68599615e-04, -1.59547254e-02, -1.42704435e-02,\n",
      "        2.08723824e-02,  2.36346312e-02,  1.94542133e-03,  1.81719160e-03,\n",
      "       -1.67929509e-03, -1.32182408e-02, -3.86173808e-04,  2.04938333e-02,\n",
      "        1.29830511e-02, -5.85814845e-03,  2.71011442e-02, -8.59373063e-03,\n",
      "        3.93978395e-02,  2.24397443e-02, -3.24693765e-03, -2.69970810e-03,\n",
      "        1.51997292e-02,  2.69879289e-02,  1.82309989e-02, -1.75221788e-03,\n",
      "        4.86637093e-02, -1.74408033e-02, -4.27516177e-03, -9.57739353e-03,\n",
      "        3.57324742e-02, -3.90069373e-03,  5.91163989e-04,  1.52269397e-02,\n",
      "        9.31583438e-03,  3.76671250e-03, -8.85934103e-03, -1.45518687e-02,\n",
      "        3.25292163e-02, -4.04688902e-03, -7.23422877e-03, -1.00737447e-02,\n",
      "        1.47045578e-03,  8.80711246e-03, -1.85284242e-02,  1.36964873e-03,\n",
      "        4.09774408e-02, -3.59211481e-05, -1.20226955e-02, -5.02190460e-03,\n",
      "        1.69506520e-02, -1.87287312e-02, -2.61590518e-02, -1.46410801e-03,\n",
      "        6.17792569e-02, -1.09061217e-02,  3.08033754e-03, -5.49047440e-03,\n",
      "       -5.25364594e-04, -1.00559806e-02, -5.79253584e-03, -6.74282480e-03,\n",
      "       -1.09802447e-02,  9.27903038e-03, -1.03342943e-02, -6.41968241e-03,\n",
      "       -6.19200757e-03, -2.71528523e-04,  2.18447447e-02,  1.16929291e-02,\n",
      "        7.32236542e-03, -1.52116064e-02, -5.73999109e-03,  4.31946628e-02,\n",
      "        1.06744478e-02,  4.33062064e-03, -2.74121924e-03, -7.20717432e-03,\n",
      "        3.02372277e-02,  2.12971158e-02,  4.18151990e-02,  3.65232071e-03,\n",
      "        3.92491110e-02, -1.76155416e-03,  9.00797267e-03,  3.58473398e-02,\n",
      "       -1.75324678e-02,  8.36529583e-03,  1.12495432e-03,  2.58151758e-02,\n",
      "       -7.42657669e-03, -2.85215117e-03, -1.17606986e-02, -2.38101128e-02,\n",
      "       -1.03344219e-02, -1.74319232e-03, -2.79709906e-03, -1.44254894e-03,\n",
      "       -3.16050206e-03, -1.81168132e-02,  6.43762806e-03,  3.88799031e-04,\n",
      "       -6.87874388e-03, -3.58888926e-03, -2.75230478e-03, -1.88389365e-02,\n",
      "        5.75838536e-02,  1.58374328e-02,  3.51316459e-03,  3.91603746e-02,\n",
      "        1.44244889e-02, -4.52182721e-03, -7.02761486e-03,  2.50094887e-02,\n",
      "       -1.65860355e-02, -9.60004423e-03,  4.06778976e-03,  4.74737346e-04,\n",
      "       -2.28861743e-03,  8.70032120e-04,  2.59555355e-02,  2.03234553e-02,\n",
      "        3.63752283e-02, -1.05150593e-02,  1.99638326e-02, -1.88530292e-02,\n",
      "       -8.81026208e-05,  1.12191895e-02, -1.20551838e-03, -1.65502876e-02,\n",
      "        4.00181971e-02, -5.10746660e-03,  2.30650343e-02,  5.88737987e-02,\n",
      "        3.36543359e-02,  1.78912804e-02,  2.70006079e-02, -1.72870525e-03,\n",
      "       -1.64663838e-03,  1.62696522e-02, -2.42688926e-04,  2.30893902e-02,\n",
      "        2.55123470e-02, -9.22810938e-03,  2.34492645e-02,  1.83226280e-02,\n",
      "       -1.00391060e-02, -8.26303940e-03, -7.36286631e-03, -1.19366217e-03,\n",
      "       -9.82848648e-03,  3.52882873e-03, -5.47762169e-03,  2.85355677e-03,\n",
      "        7.28537142e-03, -2.62638144e-02, -8.78089748e-04, -1.21890921e-02,\n",
      "        3.18433754e-02,  4.09586646e-04,  1.02058530e-03, -5.75702265e-03,\n",
      "        1.76410247e-02,  2.40592957e-02, -5.87736443e-03, -6.77918177e-03,\n",
      "        5.96781669e-04, -2.03048754e-02, -1.44697377e-03, -2.38203211e-03,\n",
      "       -4.93814144e-03,  3.15943325e-04, -6.26481650e-03, -2.20893361e-02,\n",
      "        0.00000000e+00, -1.55147328e-03, -1.58513170e-02,  8.56551994e-03,\n",
      "       -1.72020383e-02, -1.42314760e-02,  3.80211100e-02, -6.16566883e-03,\n",
      "       -2.85758893e-03,  7.19777963e-05,  3.42518911e-02,  1.90401133e-02,\n",
      "        3.43159325e-02, -1.44836232e-02,  1.91375287e-03, -5.38675673e-03,\n",
      "        3.32253873e-02,  1.14997728e-02,  4.43387516e-02, -9.00333689e-04,\n",
      "        1.47994868e-02, -6.81562629e-03,  3.70736718e-02, -1.18388282e-03,\n",
      "       -1.14756785e-02,  1.14307273e-02, -6.40728697e-03,  3.47381718e-02,\n",
      "       -8.98742117e-03, -1.11382017e-02,  6.23127446e-04, -4.87971026e-03,\n",
      "       -5.62839350e-03, -5.67940343e-03, -3.71852936e-03, -1.24859689e-02,\n",
      "        2.37574466e-02,  1.23657053e-02, -1.04385791e-02,  2.68229819e-03,\n",
      "        1.22646755e-03,  1.40425274e-02,  1.11934645e-02,  1.50690367e-03,\n",
      "        2.33984627e-02, -1.20530035e-02, -7.49909738e-03,  2.59265341e-02,\n",
      "       -4.39346954e-03, -3.66363628e-03,  4.97815839e-04, -5.19906217e-03,\n",
      "       -1.71481092e-02,  2.19373684e-02,  2.69555990e-02,  2.60442263e-04,\n",
      "        7.96250906e-03, -1.05364593e-02,  7.28427898e-04,  1.35798845e-03,\n",
      "        1.54113746e-03, -2.12011044e-03,  3.67840752e-02, -1.97012126e-02,\n",
      "       -4.94636886e-04, -4.11554473e-03,  3.00261490e-02,  6.09969124e-02,\n",
      "        3.89116630e-02,  3.18992101e-02, -7.24407239e-03,  2.43893079e-02,\n",
      "       -1.21562583e-02,  1.62003487e-02,  1.08666299e-02, -1.20254504e-02,\n",
      "        3.51043306e-02, -7.69854151e-03, -3.14192352e-04, -8.27963185e-03,\n",
      "       -6.42872765e-04, -1.28433947e-03, -1.04879374e-02, -2.27927454e-02,\n",
      "       -1.17057946e-03, -3.21979565e-03, -2.64506671e-03,  5.43384701e-02,\n",
      "        1.36379972e-02,  4.14934615e-03, -1.26534440e-02,  4.27153930e-02,\n",
      "       -3.77887953e-03, -3.24612446e-02,  9.56569798e-04,  1.00781466e-03,\n",
      "        2.58289147e-02,  4.46516164e-02, -5.99533413e-03, -1.03733093e-02,\n",
      "       -7.03210477e-03,  2.80842129e-02, -3.78092891e-03,  1.75546221e-02,\n",
      "       -2.10590549e-02, -9.91412532e-03, -4.49678209e-03, -8.53611529e-03,\n",
      "       -1.72981936e-02, -5.02727833e-03,  4.84723458e-03, -1.24454135e-02,\n",
      "       -2.19882634e-02, -4.01064707e-03,  3.00682504e-02, -1.18015166e-02,\n",
      "       -1.38439266e-02,  2.11446881e-02,  5.94557205e-04,  2.89103296e-03,\n",
      "       -1.84859633e-02, -4.34693461e-03, -1.33426702e-02, -1.34250084e-02,\n",
      "       -9.81305353e-03, -1.92418229e-02, -4.26616473e-03, -2.89671272e-02,\n",
      "       -1.73862334e-02,  3.01364996e-03,  1.82147033e-03, -1.60560478e-02,\n",
      "       -2.90597957e-02,  1.22956454e-03,  4.32112673e-03, -1.11694885e-02,\n",
      "        1.28828874e-02, -1.30875502e-04, -6.13328163e-03,  1.41407223e-02,\n",
      "       -1.07291704e-02, -6.76778005e-03, -6.99757040e-03,  1.45588424e-02,\n",
      "       -5.05492510e-03,  8.39763554e-04,  8.43331327e-06, -3.43040773e-03,\n",
      "        2.79296213e-03, -7.90322386e-03, -2.63868105e-02, -4.38705832e-03,\n",
      "        1.52690848e-02,  1.15942350e-03,  3.14513929e-02, -4.95479396e-03,\n",
      "        1.03815850e-02,  1.28943240e-03,  6.45941682e-03, -2.08903700e-02,\n",
      "        1.69702526e-02, -5.18085202e-03, -1.01624159e-02,  2.65851226e-02,\n",
      "       -5.29317139e-03, -2.24406030e-02, -3.74004710e-03, -3.87633708e-03,\n",
      "       -2.34784894e-02,  9.39205941e-03,  8.65725242e-03, -2.41578184e-03,\n",
      "        1.25047995e-03,  1.91509922e-03,  4.12717462e-03,  2.91320775e-03,\n",
      "       -8.49666260e-03,  1.96327502e-03, -4.45866404e-04,  2.88965832e-02,\n",
      "       -1.61067750e-02, -6.31900784e-03, -1.36325294e-02, -1.87324011e-03,\n",
      "       -1.20700826e-03, -1.79634895e-02, -4.65209177e-03, -7.70781562e-03,\n",
      "       -3.98072042e-03, -1.11900792e-02, -1.07051535e-02,  2.66864654e-02,\n",
      "       -1.26174456e-02, -1.30147848e-03, -2.45995726e-03, -1.13751087e-02,\n",
      "        5.81091772e-05, -2.02446105e-03,  3.82131860e-02, -2.13511568e-03,\n",
      "       -2.93310601e-02, -5.40244440e-03,  2.60777418e-02, -3.65024805e-03,\n",
      "       -1.22573413e-02,  5.34577761e-03, -9.32078063e-03,  2.71532740e-02,\n",
      "        1.79467176e-03,  2.28472892e-02, -1.30087212e-02,  9.36642569e-03,\n",
      "        8.14848021e-03,  1.12254862e-02, -3.45294061e-03, -1.19771936e-03,\n",
      "        1.22619914e-02,  3.11376601e-02, -2.10467633e-03, -5.23144519e-03],\n",
      "      dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'padding': 'valid', 'data_format': 'channels_first', 'strides': (2, 2), 'pool_size': (2, 2), 'name': 'max_pooling2d_4', 'trainable': True} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'rate': 0.2, 'name': 'dropout_4', 'seed': None, 'trainable': True, 'noise_shape': None} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'name': 'flatten_1', 'trainable': True} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'bias_regularizer': None, 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'name': 'dense_1', 'trainable': True, 'bias_constraint': None, 'activation': 'relu', 'activity_regularizer': None, 'units': 4096, 'use_bias': True, 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[ 7.5686136e-03, -7.6942574e-03, -3.1062676e-02, ...,\n",
      "         7.2872648e-03,  2.9260896e-02, -3.3054791e-02],\n",
      "       [ 1.6364675e-02, -1.1163901e-02,  9.0726856e-03, ...,\n",
      "         2.0957628e-02,  3.0666796e-02,  1.1102936e-02],\n",
      "       [-3.5514606e-03,  1.9036731e-02, -2.2466309e-02, ...,\n",
      "         1.4672839e-02,  3.0978888e-02,  9.6692387e-03],\n",
      "       ...,\n",
      "       [-1.1971448e-02, -1.4063584e-02,  3.3996258e-02, ...,\n",
      "         8.0964863e-03, -3.2749642e-02,  2.1900958e-05],\n",
      "       [-3.3654545e-02, -1.6173467e-02, -2.0401921e-02, ...,\n",
      "         1.8719995e-02, -3.1253986e-02, -2.0209638e-02],\n",
      "       [-3.0314870e-02, -3.6352763e-03,  1.0174350e-02, ...,\n",
      "         3.3698194e-02,  1.2361268e-02, -2.8556254e-02]], dtype=float32), array([-0.01204074,  0.00042627, -0.00549379, ..., -0.00213445,\n",
      "        0.01065413, -0.00259949], dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'rate': 0.2, 'name': 'dropout_5', 'seed': None, 'trainable': True, 'noise_shape': None} \n",
      "\n",
      "h==  [] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n",
      "g==  {'bias_regularizer': None, 'kernel_initializer': {'config': {'mode': 'fan_avg', 'seed': None, 'scale': 1.0, 'distribution': 'uniform'}, 'class_name': 'VarianceScaling'}, 'bias_initializer': {'config': {}, 'class_name': 'Zeros'}, 'name': 'dense_2', 'trainable': True, 'bias_constraint': None, 'activation': 'softmax', 'activity_regularizer': None, 'units': 9, 'use_bias': True, 'kernel_regularizer': None, 'kernel_constraint': None} \n",
      "\n",
      "h==  [array([[-0.00247621, -0.02238109,  0.02638318, ...,  0.01242757,\n",
      "         0.03426337,  0.01604994],\n",
      "       [ 0.03622317, -0.02517531,  0.0222188 , ...,  0.0169722 ,\n",
      "         0.01126354,  0.03779897],\n",
      "       [ 0.01498923, -0.01751323, -0.02430623, ..., -0.02996848,\n",
      "        -0.02401111, -0.01065238],\n",
      "       ...,\n",
      "       [ 0.0266019 ,  0.02426228, -0.00161164, ..., -0.03282434,\n",
      "         0.02428572,  0.00285495],\n",
      "       [-0.00837479,  0.01713976,  0.02187984, ..., -0.0162278 ,\n",
      "         0.00307601, -0.01597011],\n",
      "       [-0.02792199, -0.00302457, -0.00782866, ..., -0.01805228,\n",
      "         0.03802843, -0.01716576]], dtype=float32), array([-0.43791878,  0.08543716,  0.14084587,  0.28243527,  0.19612001,\n",
      "        0.40762126, -0.17842455, -0.19597583, -0.30014047], dtype=float32)] \n",
      "\n",
      "\n",
      "type_of g ==  <class 'dict'> \n",
      "\n",
      "type_of h ==  <class 'list'> \n",
      "\n"
     ]
    }
   ],
   "source": [
    "from keras.models import load_model\n",
    "\n",
    "model = load_model(learning_model_path+'learning_model.h5')\n",
    "#model = load_model('/home/atif/traffic_model_11_dec_1_filter.h5')\n",
    "layer_list =[]\n",
    "# f = open(learning_model_path+'model_file_info.txt', 'w') #uncomment it if you want to store all layer info at a time.\n",
    "\n",
    "def make_file(g1=0,h1=0,g_type=0,h_type=0, path=0, file_name=0):\n",
    "    f = open(path + file_name+'.txt','a')  # uncomment it if you want to store all layer info at a time.\n",
    "    f.write(\"layer_definition: \" + g1 + \"\\n\\n\")\n",
    "    f.write(\"layer_type: \" + g_type + \"\\n\\n\")\n",
    "    # f.write(\"\\n\")\n",
    "    f.write(\"layer_weight: \" + h1 + \"\\n\\n\")\n",
    "    f.write(\"weight_type: \" + h_type + \"\\n\\n\\n\")\n",
    "    f.write(\"\\n\")\n",
    "    f.close()\n",
    "\n",
    "\n",
    "\n",
    "for layer in model.layers:\n",
    "    g=layer.get_config()\n",
    "    h=layer.get_weights()\n",
    "    \n",
    "    layer_list.append(h)\n",
    "    print (\"g== \",g,\"\\n\") #for printing layer name and verbal info\n",
    "\n",
    "    print (\"h== \",h,\"\\n\\n\") # for printing layer numeric value, eg: weight, bias value\n",
    "    print(\"type_of g == \",type(g),\"\\n\")\n",
    "    print(\"type_of h == \",type(h),\"\\n\")\n",
    "\n",
    "# below lines till f.close() used for writing in text file. To do this you have to uncomment the above line started with f.open() also.\n",
    "\n",
    "    g1=str(g) # declaring a string variable g1 to store the info of g\n",
    "    h1=str(h) #declaring a string variable h1 to store the info of h\n",
    "    g_type=str(type(g)) #declaring a string variable g1 to store the type of g\n",
    "    h_type=str(type(h)) #declaring a string variable h1 to store the type of h\n",
    "    make_file(g1, h1, g_type, h_type, learning_model_path, file_name=\"abc\")\n",
    "    # f.write(\"layer_definition: \"+g1+\"\\n\\n\")\n",
    "    # f.write(\"layer_type: \"+g_type+\"\\n\\n\")\n",
    "    # #f.write(\"\\n\")\n",
    "    # f.write(\"layer_weight: \"+h1+\"\\n\\n\")\n",
    "    # f.write(\"weight_type: \"+h_type+\"\\n\\n\\n\")\n",
    "    # f.write(\"\\n\")\n",
    "    \n",
    "f.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(layer_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "conv_kernel: \n",
      " [[[[-3.09373170e-01 -1.42402902e-01 -2.26866543e-01]\n",
      "   [-3.38641018e-01 -1.47521928e-01 -3.19729000e-01]\n",
      "   [-2.69378066e-01 -1.13230884e-01 -1.55600101e-01]]]\n",
      "\n",
      "\n",
      " [[[ 3.35247469e+00  3.27248263e+00  3.39521170e+00]\n",
      "   [ 3.43956995e+00  3.48220563e+00  3.13530016e+00]\n",
      "   [ 3.29916310e+00  3.28517079e+00  3.32652211e+00]]]\n",
      "\n",
      "\n",
      " [[[-8.17258190e-03  4.66080382e-02 -6.72599450e-02]\n",
      "   [-4.37568873e-02 -1.10951774e-02 -1.23586863e-01]\n",
      "   [ 1.00698866e-01  1.48765622e-02  6.54247105e-02]]]\n",
      "\n",
      "\n",
      " [[[-1.52699903e-01 -5.47038130e-02  4.48007286e-02]\n",
      "   [ 2.05426086e-02 -3.40171345e-02 -5.17112538e-02]\n",
      "   [-1.02891922e-02  6.80713132e-02 -1.51523337e-01]]]\n",
      "\n",
      "\n",
      " [[[-3.92200202e-02 -1.53161272e-01 -1.82539582e-01]\n",
      "   [-6.38462678e-02 -8.27020258e-02  4.74050529e-02]\n",
      "   [ 2.16074809e-02 -8.74133501e-03  8.50881413e-02]]]\n",
      "\n",
      "\n",
      " [[[-4.09431569e-02 -5.82188144e-02  1.98313389e-02]\n",
      "   [-8.28460529e-02  2.29802541e-02 -5.44978157e-02]\n",
      "   [-9.02320668e-02  3.28972042e-02  3.04461997e-02]]]\n",
      "\n",
      "\n",
      " [[[-6.50194585e-02  7.87916705e-02  1.07447848e-01]\n",
      "   [ 2.52780113e-02 -1.33519992e-01 -9.57548171e-02]\n",
      "   [-1.16692565e-01  1.30822301e-01 -8.85324851e-02]]]\n",
      "\n",
      "\n",
      " [[[ 3.95465112e+00  3.95243692e+00  3.99482274e+00]\n",
      "   [ 4.17110682e+00  4.10947704e+00  3.91495299e+00]\n",
      "   [ 4.02014494e+00  4.08517122e+00  3.89690208e+00]]]\n",
      "\n",
      "\n",
      " [[[-2.88339257e-02 -8.48340243e-03  4.31856215e-02]\n",
      "   [-1.41278327e-01 -7.26879463e-02 -1.18388310e-01]\n",
      "   [ 3.62323821e-02 -2.98831090e-02 -3.58887692e-03]]]\n",
      "\n",
      "\n",
      " [[[ 9.65168417e-01  1.05669081e+00  8.05238843e-01]\n",
      "   [ 8.75337839e-01  1.03724945e+00  8.12726080e-01]\n",
      "   [ 9.14092720e-01  9.31659162e-01  8.93082559e-01]]]\n",
      "\n",
      "\n",
      " [[[ 2.26394725e+00  2.10688400e+00  2.12024355e+00]\n",
      "   [ 2.24116945e+00  2.14315200e+00  2.09675884e+00]\n",
      "   [ 2.20846891e+00  2.30330276e+00  2.20282936e+00]]]\n",
      "\n",
      "\n",
      " [[[ 9.19752643e-02 -1.34589359e-01 -3.20941322e-02]\n",
      "   [-5.90169020e-02  8.78517702e-02 -6.33397698e-02]\n",
      "   [-8.22363496e-02 -1.13465577e-01 -2.61048358e-02]]]\n",
      "\n",
      "\n",
      " [[[ 1.04308200e+00  8.76845360e-01  9.62297499e-01]\n",
      "   [ 1.12359333e+00  1.07152927e+00  9.58534420e-01]\n",
      "   [ 9.95436549e-01  1.04079044e+00  9.17856395e-01]]]\n",
      "\n",
      "\n",
      " [[[ 5.02101851e+00  4.82489967e+00  4.94219542e+00]\n",
      "   [ 4.99869680e+00  5.00246525e+00  4.83660984e+00]\n",
      "   [ 4.87846518e+00  4.97926092e+00  4.70491791e+00]]]\n",
      "\n",
      "\n",
      " [[[-1.16733693e-01  5.96768111e-02 -9.95145589e-02]\n",
      "   [-1.12865284e-01 -8.25458243e-02 -1.39002670e-02]\n",
      "   [-1.24915928e-01  4.46644910e-02 -7.18671605e-02]]]\n",
      "\n",
      "\n",
      " [[[-6.56417534e-02  4.66563962e-02 -1.39926538e-01]\n",
      "   [-1.11755384e-02 -1.28494158e-01  8.78791735e-02]\n",
      "   [ 9.53978822e-02 -6.95486292e-02  5.12253679e-02]]]\n",
      "\n",
      "\n",
      " [[[ 2.01192570e+00  1.94212937e+00  1.90619969e+00]\n",
      "   [ 1.86557484e+00  1.76456392e+00  1.95243657e+00]\n",
      "   [ 1.73148465e+00  1.81070042e+00  1.91491842e+00]]]\n",
      "\n",
      "\n",
      " [[[ 9.72868279e-02  1.14101104e-01 -9.37294587e-02]\n",
      "   [-2.17131153e-02 -1.22840561e-01 -2.67206207e-02]\n",
      "   [-9.55658965e-03  2.16994155e-02 -1.08022571e-01]]]\n",
      "\n",
      "\n",
      " [[[-1.16467252e-01  4.62026633e-02  5.46604730e-02]\n",
      "   [ 1.07350290e-01 -9.87263620e-02 -4.49119024e-02]\n",
      "   [ 1.82193909e-02 -1.30475745e-01 -1.18728414e-01]]]\n",
      "\n",
      "\n",
      " [[[ 1.52933329e-01  3.23026255e-02 -2.17119232e-02]\n",
      "   [-2.62478907e-02 -5.86965261e-03 -3.40639129e-02]\n",
      "   [ 4.85175289e-02 -6.84672445e-02 -7.95996934e-02]]]\n",
      "\n",
      "\n",
      " [[[ 1.46493137e-01  1.27386332e-01 -1.29598424e-01]\n",
      "   [-1.33075669e-01 -1.00361601e-01 -1.30016267e-01]\n",
      "   [ 3.25670429e-02  2.04589795e-02 -1.39560878e-01]]]\n",
      "\n",
      "\n",
      " [[[ 2.17062030e-02 -1.30084738e-01  9.41729695e-02]\n",
      "   [ 1.76684577e-02 -3.60378958e-02 -8.57137293e-02]\n",
      "   [ 1.53056430e-02 -1.00446366e-01 -1.25697985e-01]]]\n",
      "\n",
      "\n",
      " [[[ 5.51404692e-02  3.49206328e-02 -9.99556780e-02]\n",
      "   [-1.00192830e-01 -3.16587724e-02 -8.34661499e-02]\n",
      "   [ 9.67345014e-02 -4.77962270e-02 -7.29252994e-02]]]\n",
      "\n",
      "\n",
      " [[[ 3.53696299e+00  3.36314869e+00  3.43063521e+00]\n",
      "   [ 3.45771503e+00  3.29109192e+00  3.37403655e+00]\n",
      "   [ 3.40354753e+00  3.21759295e+00  3.31700110e+00]]]\n",
      "\n",
      "\n",
      " [[[ 4.58308786e-01  5.77509284e-01  2.99992830e-01]\n",
      "   [ 5.16355932e-01  3.97692233e-01  5.54753780e-01]\n",
      "   [ 4.47713226e-01  3.13956141e-01  4.51437563e-01]]]\n",
      "\n",
      "\n",
      " [[[-1.42728373e-01 -2.66903043e-02 -1.25056831e-03]\n",
      "   [-2.03536615e-01 -1.55687690e-01 -5.16405441e-02]\n",
      "   [-1.65807560e-01  3.22208442e-02 -1.54678985e-01]]]\n",
      "\n",
      "\n",
      " [[[-4.09316383e-02 -2.27216091e-02 -1.38971200e-02]\n",
      "   [-1.34889647e-01  7.56319240e-02 -9.30356421e-03]\n",
      "   [-2.48249155e-02 -1.23100474e-01  1.51565550e-02]]]\n",
      "\n",
      "\n",
      " [[[ 3.49415421e+00  3.47769499e+00  3.28418851e+00]\n",
      "   [ 3.60002398e+00  3.53860712e+00  3.39549541e+00]\n",
      "   [ 3.39432311e+00  3.35459828e+00  3.43135619e+00]]]\n",
      "\n",
      "\n",
      " [[[-9.86798704e-02 -4.15516235e-02  1.11651063e-01]\n",
      "   [-1.41094133e-01  2.89036869e-03  1.07618041e-01]\n",
      "   [-1.05185686e-02 -2.12714523e-02 -1.08369745e-01]]]\n",
      "\n",
      "\n",
      " [[[-7.59547353e-02  7.84290880e-02 -1.16201900e-01]\n",
      "   [-1.30550429e-01  7.64918998e-02 -1.24180317e-01]\n",
      "   [-3.40212919e-02  8.00873414e-02 -3.11297365e-02]]]\n",
      "\n",
      "\n",
      " [[[ 3.87289023e+00  3.85823417e+00  3.64895511e+00]\n",
      "   [ 3.80404592e+00  3.74333215e+00  3.73351955e+00]\n",
      "   [ 3.68116856e+00  3.55919170e+00  3.55899215e+00]]]\n",
      "\n",
      "\n",
      " [[[ 2.35459709e+00  2.40642476e+00  2.14763641e+00]\n",
      "   [ 2.31310749e+00  2.19153237e+00  2.28144336e+00]\n",
      "   [ 2.26339674e+00  2.13463354e+00  2.09235477e+00]]]] \n",
      "\n",
      "\n"
     ]
    },
    {
     "ename": "IndexError",
     "evalue": "list index out of range",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-6-2601286adc9c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     16\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 18\u001b[0;31m \u001b[0mdense_kernel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlayer_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     19\u001b[0m \u001b[0;31m#print(\"dense_kernel: \\n\",dense_kernel,\"\\n\\n\")\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     20\u001b[0m \u001b[0;31m#print(\"dense_kernel shape:\\t\",dense_kernel.shape,\"\\n\\n\")\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mIndexError\u001b[0m: list index out of range"
     ]
    }
   ],
   "source": [
    "#From here the code has started which will extract every layer's info which you can use further in this file\n",
    "layer_list\n",
    "conv_kernel=layer_list[0][0]\n",
    "conv_kernel=conv_kernel.transpose()\n",
    "print(\"conv_kernel: \\n\",conv_kernel,\"\\n\\n\")\n",
    "# print(\"conv_kernel shape:\\t\",conv_kernel.shape,\"\\n\\n\")\n",
    "# print(\"conv kernel dimension:\\t\",conv_kernel.ndim,\"\\n\\n\")\n",
    "# print(\"type_conv_kernel:\",type(conv_kernel),\"\\n\")\n",
    "\n",
    "\n",
    "\n",
    "conv_bias=layer_list[0][1]\n",
    "# print(\"conv_bias_value: \",conv_bias)\n",
    "# print(\"conv_bias ndim: \",conv_bias.ndim,\"\\n\\n\")\n",
    "\n",
    "\n",
    "\n",
    "dense_kernel=layer_list[2][0]\n",
    "#print(\"dense_kernel: \\n\",dense_kernel,\"\\n\\n\")\n",
    "#print(\"dense_kernel shape:\\t\",dense_kernel.shape,\"\\n\\n\")\n",
    "#print(\"dense_kernel dimension:\\t\",dense_kernel.ndim,\"\\n\\n\")\n",
    "#print(\"type_dense_kernel:\",type(dense_kernel),\"\\n\")\n",
    "#print(\"dense_kernel size: \",dense_kernel.size,\"\\n\")\n",
    "# dense_1_transpose=dense__1.transpose()\n",
    "# print(\"dense_1_transpose: \",dense_1_transpose,\"\\n\\n\")\n",
    "\n",
    "\n",
    "dense_bias=layer_list[2][1]\n",
    "#print(\"dense_bias: \",dense_bias)\n",
    "#print(\"dense_bias_shape: \",dense_bias.shape)\n",
    "dense_bias=dense_bias.reshape(1,NUM_CLASSES) # here chenge 5 to the number of your used class\n",
    "#print(\"dense_bias_shape: \",dense_bias.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([[[-0.67074907, -0.59329855, -0.53431404],\n",
      "        [-0.7462511 , -0.3903124 ,  0.06319907],\n",
      "        [-0.36227798, -0.09837703,  0.5340042 ]]], dtype=float32), array([[[-0.26519078, -0.19100626,  0.2422876 ],\n",
      "        [-0.20440044, -0.29235226,  0.07251699],\n",
      "        [ 0.01618635, -0.33096606,  0.04459456]]], dtype=float32), array([[[-1.5925051 , -0.5391095 , -0.56155765],\n",
      "        [-0.5556934 ,  0.7204196 ,  1.0512314 ],\n",
      "        [-0.18157953,  1.145218  ,  1.6162105 ]]], dtype=float32), array([[[ 0.0586549 ,  0.04417358, -0.21519895],\n",
      "        [-0.09233072, -0.32553932,  0.0979021 ],\n",
      "        [ 0.238696  , -0.09672035,  0.1477777 ]]], dtype=float32), array([[[ 0.43710935, -0.6779395 , -1.9064794 ],\n",
      "        [ 1.566656  ,  0.9854658 , -0.7540373 ],\n",
      "        [ 1.5616229 ,  0.89573854, -0.4884904 ]]], dtype=float32)]\n",
      "[[[[-0.67074907 -0.59329855 -0.53431404]\n",
      "   [-0.7462511  -0.3903124   0.06319907]\n",
      "   [-0.36227798 -0.09837703  0.5340042 ]]]\n",
      "\n",
      "\n",
      " [[[-0.26519078 -0.19100626  0.2422876 ]\n",
      "   [-0.20440044 -0.29235226  0.07251699]\n",
      "   [ 0.01618635 -0.33096606  0.04459456]]]\n",
      "\n",
      "\n",
      " [[[-1.5925051  -0.5391095  -0.56155765]\n",
      "   [-0.5556934   0.7204196   1.0512314 ]\n",
      "   [-0.18157953  1.145218    1.6162105 ]]]\n",
      "\n",
      "\n",
      " [[[ 0.0586549   0.04417358 -0.21519895]\n",
      "   [-0.09233072 -0.32553932  0.0979021 ]\n",
      "   [ 0.238696   -0.09672035  0.1477777 ]]]\n",
      "\n",
      "\n",
      " [[[ 0.43710935 -0.6779395  -1.9064794 ]\n",
      "   [ 1.566656    0.9854658  -0.7540373 ]\n",
      "   [ 1.5616229   0.89573854 -0.4884904 ]]]]\n",
      "4\n",
      "[ 0.63377595 -0.01135568 -0.02604145 -0.11364901 -0.0088685 ]\n",
      "convolution_kernel_filter: \n",
      " [[[-0.67074907 -0.59329855 -0.53431404]\n",
      "  [-0.74625111 -0.3903124   0.06319907]\n",
      "  [-0.36227798 -0.09837703  0.53400421]]\n",
      "\n",
      " [[-0.26519078 -0.19100626  0.24228761]\n",
      "  [-0.20440044 -0.29235226  0.07251699]\n",
      "  [ 0.01618635 -0.33096606  0.04459456]]\n",
      "\n",
      " [[-1.5925051  -0.53910953 -0.56155765]\n",
      "  [-0.55569339  0.72041959  1.05123138]\n",
      "  [-0.18157953  1.14521801  1.61621046]]\n",
      "\n",
      " [[ 0.0586549   0.04417358 -0.21519895]\n",
      "  [-0.09233072 -0.32553932  0.0979021 ]\n",
      "  [ 0.23869599 -0.09672035  0.14777771]]\n",
      "\n",
      " [[ 0.43710935 -0.67793947 -1.90647936]\n",
      "  [ 1.56665599  0.98546582 -0.75403732]\n",
      "  [ 1.56162286  0.89573854 -0.4884904 ]]] \n",
      "\n",
      "convolution_kernel_filter shape: (5, 3, 3) \tconvolution_kernel_filter ndim: 3 \n",
      "\n",
      "length of convolution_kernel_filter: 5 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "conv_kernel=layer_list[0][0]\n",
    "conv_kernel=conv_kernel.transpose() # This has made to print it like a Matrix form. \n",
    "conv_kernel_list=[]\n",
    "for i in conv_kernel:\n",
    "#     print(i)\n",
    "    conv_kernel_list.append(i)\n",
    "#     for k in i:\n",
    "#         print(k)\n",
    "#         i_list.append(k)\n",
    "print(conv_kernel_list)\n",
    "conv_kernel_list_array=[]\n",
    "conv_kernel_list_array=np.array(conv_kernel_list)\n",
    "print(conv_kernel_list_array)\n",
    "# i_list_array=i_list_array.reshape(2,3,3)\n",
    "print(conv_kernel_list_array.ndim)\n",
    "\n",
    "\n",
    "# look conv_kernel_list_array[x][y][z][w] -- it is a 4 dim array. x is number of filter, y is full stack, z is number of row of each filter, w is each single element of every row\n",
    "for i in range(number_filter):\n",
    "    for j in range(filter_size):\n",
    "        for k in range(filter_size):\n",
    "            p = conv_kernel_list_array[i][0][j][k]\n",
    "            ww = str(p)\n",
    "            ww = ww.replace('[', '')\n",
    "            ww = ww.replace(']', '')\n",
    "            # ww = ww.strip()\n",
    "            f = open(learning_model_path + 'conv_kernel.txt','a')  # uncomment from here till f.close() if you want to save text file\n",
    "            f.write(ww)\n",
    "            if k<(filter_size-1):\n",
    "                f.write(\" \")\n",
    "        f.write(\"\\n\")\n",
    "f.close()\n",
    "\n",
    "\n",
    "# =============================================================================\n",
    "# Storing convolution bias, needed for cpp testing\n",
    "# =============================================================================\n",
    "\n",
    "conv_bias=layer_list[0][1]\n",
    "conv_bias_list=[]\n",
    "for i in conv_bias:\n",
    "    conv_bias_list.append(i)\n",
    "\n",
    "conv_bias_array=[]\n",
    "conv_bias_array=np.array(conv_bias_list)\n",
    "print(conv_bias_array)\n",
    "np.savetxt(learning_model_path+'conv_bias.txt', conv_bias_array, fmt='%1.8e',delimiter=' ')\n",
    "\n",
    "#for x in conv_bias_array:\n",
    "    #print(x)\n",
    "    \n",
    "    \n",
    "\n",
    "# =============================================================================\n",
    "# Storing dense kernel weight, needed for cpp testing\n",
    "# =============================================================================\n",
    " \n",
    "dense_kernel=layer_list[2][0]\n",
    "i_list=[] #declare a list to store the weight of dense kernel\n",
    "for i in dense_kernel:\n",
    "#     print(i)\n",
    "    i_list.append(i) #appended it in the declared list\n",
    "#     for k in i:\n",
    "# #         print(k)\n",
    "#         i_list.append(k)\n",
    "# print(i_list)\n",
    "i_list_array=[] #declared an array\n",
    "i_list_array=np.array(i_list) # store the value of list in the array\n",
    "#print(i_list_array)\n",
    "np.savetxt(learning_model_path+'dense_kernel.txt', i_list_array, fmt='%1.8e',delimiter=' ') #writing on a text file from array\n",
    "\n",
    "# %.8f #you can use it to get float value\n",
    "# fmt='%1.8e' #add this above line after i_list_aray\n",
    "\n",
    "# =============================================================================\n",
    "# Storing dense bias, needed for cpp testing\n",
    "# =============================================================================\n",
    "\n",
    "dense_bias=layer_list[2][1]\n",
    "dense_bias_list=[]\n",
    "for i in dense_bias:\n",
    "    dense_bias_list.append(i)\n",
    "\n",
    "dense_bias_array=[]\n",
    "dense_bias_array=np.array(dense_bias_list)\n",
    "#print(dense_bias_array)\n",
    "np.savetxt(learning_model_path+'dense_bias.txt', dense_bias_array, fmt='%1.8e',delimiter=' ')\n",
    "\n",
    "\n",
    "# =============================================================================\n",
    "# Reshaping convolution kernel for further process\n",
    "# =============================================================================\n",
    "\n",
    "# print(\"conv_kernel:\\n\",conv_kernel,\"\\n\")\n",
    "# print(\"conv_kernel_shape:\",conv_kernel.shape,\"\\tconv_kernel ndim:\",conv_kernel.ndim,\"\\n\")\n",
    "# print(\"length of conv_kernel:\",len(conv_kernel),\"\\n\")\n",
    "\n",
    "conv_kernel_reshape=conv_kernel.reshape(number_filter,3,3) # 2 for 2 filter. change it according to your filter number\n",
    "# print(\"conv_kernel_reshape:\\n\",conv_kernel_reshape,\"\\n\")\n",
    "# print(\"conv_kernel_reshape shape:\",conv_kernel_reshape.shape,\"\\tconv_kernel_reshape ndim:\",conv_kernel_reshape.ndim,\"\\n\")\n",
    "# print(\"length of conv_kernel_reshape:\",len(conv_kernel_reshape[0]),\"\\n\")\n",
    "\n",
    "convolution_kernel_filter=[]\n",
    "convolution_kernel_filter=np.zeros((number_filter,3,3)) # 2 for 2 filter. change it according to your filter number\n",
    "convolution_kernel_filter[:,:,:]=np.array(conv_kernel_reshape)\n",
    "print(\"convolution_kernel_filter: \\n\",convolution_kernel_filter,\"\\n\")\n",
    "print(\"convolution_kernel_filter shape:\",convolution_kernel_filter.shape,\"\\tconvolution_kernel_filter ndim:\",convolution_kernel_filter.ndim,\"\\n\")\n",
    "print(\"length of convolution_kernel_filter:\",len(convolution_kernel_filter),\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [],
   "source": [
    "def conv_(img, conv_filter, conv_bias_val, number_filter):\n",
    "#     print(\"\\nconv_ function start to work\\n\")\n",
    "#     print(\"conv_filter: \",conv_filter)\n",
    "#     print(\"eee: \",conv_bias_val, \"\\t num_filter: \", number_filter)\n",
    "    filter_size = conv_filter.shape[1] #output is 3\n",
    "#    print(\"filter_size: \",filter_size)\n",
    "#    print(\"img shape: \",img.shape)\n",
    "    result = np.zeros((img.shape))\n",
    "#    Looping through the image to apply the convolution operation.\n",
    "    \n",
    "    for x in range(number_filter):  # this loop take decision looping will be continued to num_filter times. Why?? because then it will take one by one \n",
    "                                            # all filter and perform the convolution.\n",
    "        \n",
    "        for r in np.uint16(np.arange(filter_size/2.0,img.shape[0]-filter_size/2.0+1)):\n",
    "        \n",
    "            for c in np.uint16(np.arange(filter_size/2.0,img.shape[1]-filter_size/2.0+1)):\n",
    "            \n",
    "                curr_region = img[r-np.uint16(np.floor(filter_size/2.0)):r+np.uint16(np.ceil(filter_size/2.0)), c-np.uint16(np.floor(filter_size/2.0)):c+np.uint16(np.ceil(filter_size/2.0))]\n",
    "        \n",
    "                #Element-wise multipliplication between the current region and the filter.\n",
    "            \n",
    "                curr_result = curr_region * conv_filter\n",
    "#                 print(\"curr_result: \",curr_result)\n",
    "#                 curr_result= curr_result+x\n",
    "                \n",
    "#             print(\"conv_bias: \",conv_bias_new)\n",
    "#                 print(\"new curr res: \",curr_result)\n",
    "                conv_sum = np.sum(curr_result) #Summing the result of multiplication.\n",
    "                result[r, c] = conv_sum#Saving the summation in the convolution layer feature map.\n",
    "#        print(\"now x is: \",x)\n",
    "        #print(curr_region)\n",
    "    #Clipping the outliers of the result matrix.\n",
    "#     print(\"conv_sum_shape: \",conv_sum.shape)\n",
    "    final_result = result[np.uint16(filter_size/2.0):result.shape[0]-np.uint16(filter_size/2.0),np.uint16(filter_size/2.0):result.shape[1]-np.uint16(filter_size/2.0)]\n",
    "#     print(\"bias: \", conv_bias_val)\n",
    "    final_result = final_result+conv_bias_val\n",
    "#     print(\"\\nconv_ function finish\\n\")\n",
    "    return final_result\n",
    "\n",
    "def conv(img, conv_filter, conv_bias_array, number_filter):\n",
    "    print(\"in func: \",img.shape,\"\\t\",conv_filter.shape)\n",
    "#    print(\"\\nconv function start to work\")\n",
    "    if len(img.shape) > 2 or len(conv_filter.shape) > 3: # Check if number of image channels matches the filter depth.\n",
    "        if img.shape[-1] != conv_filter.shape[-1]:\n",
    "            print(\"Error: Number of channels in both image and filter must match.\")\n",
    "            sys.exit()\n",
    "    if conv_filter.shape[1] != conv_filter.shape[2]: # Check if filter dimensions are equal.\n",
    "        print('Error: Filter must be a square matrix. I.e. number of rows and columns must match.')\n",
    "        sys.exit()\n",
    "    if conv_filter.shape[1]%2==0: # Check if filter diemnsions are odd.\n",
    "        print('Error: Filter must have an odd size. I.e. number of rows and columns must be odd.')\n",
    "        sys.exit()\n",
    "\n",
    "    # An empty feature map to hold the output of convolving the filter(s) with the image.\n",
    "    feature_maps = np.zeros((img.shape[0]-conv_filter.shape[1]+1, img.shape[1]-conv_filter.shape[1]+1, conv_filter.shape[0]))\n",
    "\n",
    "    # Convolving the image by the filter(s).\n",
    "    \n",
    "    for filter_num in range(conv_filter.shape[0]): # number_of_filter times looping will be occurred \n",
    "#         print(\"filter num: \",filter_num)\n",
    "#        print(\"Filter \", filter_num + 1)\n",
    "        curr_filter = conv_filter[filter_num, :] # getting a filter from the bank. In every looping new filter will come here\n",
    "#         print(\"curr_filter: \",curr_filter)\n",
    "#         print(\"curr_fiter_shape: \",curr_filter.shape)\n",
    "#         print(\"length of curr_fiter_shape: \",len(curr_filter.shape))\n",
    "\n",
    "\n",
    "        if len(curr_filter.shape) > 2: # RGB filter/ Used for training RGB images\n",
    "            conv_map = conv_(img[:, :, 0], curr_filter[:, :, 0]) # Array holding the sum of all feature maps.\n",
    "            for ch_num in range(1, curr_filter.shape[-1]): # Convolving each channel with the image and summing the results.\n",
    "                conv_map = conv_map + conv_(img[:, :, ch_num], \n",
    "                                  curr_filter[:, :, ch_num])\n",
    "        else: # There is just a single channel in the filter. Black and White\n",
    "#            print(\"\\nGo to conv_ function \")\n",
    "            bias_conv = conv_bias_array[filter_num]\n",
    "#             print(\"bias_conv: \",bias_conv)\n",
    "            conv_map = conv_(img, curr_filter, bias_conv, number_filter)\n",
    "        \n",
    "        \n",
    "#         print(\"conv_map shape: \", conv_map.shape)\n",
    "            \n",
    "        feature_maps[:, :, filter_num] = conv_map # Holding feature map with the current filter.\n",
    "#         print(\"feature_maps from conv_map: \",feature_maps)\n",
    "    return feature_maps # Returning all feature maps.\n",
    "\n",
    "\n",
    "def relu(feature_map):\n",
    "    #Preparing the output of the ReLU activation function.\n",
    "    relu_out = np.zeros(feature_map.shape)\n",
    "    for map_num in range(feature_map.shape[-1]):\n",
    "        for r in np.arange(0,feature_map.shape[0]):\n",
    "            for c in np.arange(0, feature_map.shape[1]):\n",
    "                relu_out[r, c, map_num] = np.max([feature_map[r, c, map_num], 0])\n",
    "    return relu_out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = test_image_path\n",
    "# path = r'/home/atif/training_by_several_learning_process/number_classify/rgb_2_gray/Image-classification/test_image/'\n",
    "#for image in img_path:\n",
    "img_path = glob.glob(path+ '/*'+str(img_type))\n",
    "\n",
    "def do_math(img_path, convolution_kernel_filter,conv_bias_array, IMG_SIZE,number_filter, NUM_CLASSES, total_time,dense_kernel, dense_bias):\n",
    "    for image_number, image in enumerate(img_path):\n",
    "        #    print(\"\\nnum of image is: \",image_number) #It will return current image number. But careful it's cunting starts from zero so don't forget to add 1\n",
    "        print(\"\\nName of loaded image: \", image)\n",
    "        # =============================================================================\n",
    "        #     X_test=[]\n",
    "        #     X_test.append(preprocess_img(io.imread(image)))\n",
    "        #     X_test = np.array(X_test)\n",
    "        #     print(\"\\nshape: \",X_test.shape)\n",
    "        #     X_test = X_test.reshape(IMG_SIZE,IMG_SIZE)\n",
    "        #     plt.imshow(X_test)\n",
    "        # =============================================================================\n",
    "\n",
    "        X_test = []\n",
    "        #    cv_img = []\n",
    "        n = cv2.imread(image)\n",
    "        n = cv2.resize(n, (IMG_SIZE, IMG_SIZE))\n",
    "        n = cv2.cvtColor(n, cv2.COLOR_RGB2GRAY)\n",
    "        #    n = cv2.normalize(n, n, 0, 255, cv2.NORM_MINMAX)\n",
    "        n = cv2.normalize(n, None, alpha=0, beta=1, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_64F)\n",
    "\n",
    "        X_test.append(n)\n",
    "        X_test = np.array(X_test)\n",
    "        X_test = X_test.reshape(IMG_SIZE, IMG_SIZE)\n",
    "        X_test = np.rollaxis(X_test, -1)\n",
    "        print(\"img shape: \",X_test.shape,\" len fo shape: \",len(X_test.shape),\" ndim: \",X_test.ndim)\n",
    "        print(\"conv kernel shape: \",convolution_kernel_filter.shape,\" len fo shape: \",len(convolution_kernel_filter.shape),\" ndim: \",convolution_kernel_filter.ndim)\n",
    "\n",
    "        start = time.time()\n",
    "\n",
    "        feature = conv(img=X_test, conv_filter=convolution_kernel_filter, conv_bias_array=conv_bias_array, number_filter=number_filter)  # conv function calling\n",
    "        \n",
    "        print(\"feature_shape: \",feature.shape,\" feature ndim: \",feature.ndim)\n",
    "        \n",
    "#         print(\"---------\\n\",feature)\n",
    "        relu_out = relu(feature)  # relu function calling\n",
    "\n",
    "        # output of feature map / conv function\n",
    "\n",
    "        #    print(\"\\nfeature shape: \\n\",feature.shape)\n",
    "\n",
    "        #     x_feature_map=np.flipud(feature[0])\n",
    "\n",
    "        transpose_feature_map = feature.transpose()\n",
    "\n",
    "        #    print(\"\\ntranspose_feature_map shape: \",transpose_feature_map.shape)\n",
    "        #     plt.imshow(transpose_feature_map[0])\n",
    "        #    print(\"\\ntranspose_feature_map: \\n\",transpose_feature_map)\n",
    "\n",
    "        #    print(\"\\nrelu_out shape: \",relu_out.shape)\n",
    "        relu_out_transpose = relu_out.transpose()\n",
    "        print(\"\\nrelu_out_transpose shape: \",relu_out_transpose.shape)\n",
    "#         print(\"---------\\n\",relu_out_transpose)\n",
    "        #     plt.imshow(relu_out_transpose[0])\n",
    "        #    print(\"\\nrelu_out_transpose:\\n\",relu_out_transpose)\n",
    "\n",
    "        # =============================================================================\n",
    "        #      matrix multiplication with dense kernel and relu o/p\n",
    "        # =============================================================================\n",
    "\n",
    "        flatten_relu_out_transpose = relu_out_transpose.reshape(1,number_filter * 46 * 46)  # if you don't do padd on input image please make it 46*46. how 46 came?\n",
    "        # the formula of output size. and 2 for 2 filter\n",
    "#         print(\"\\nflatten_relu_out_transpose shape: \\n\",flatten_relu_out_transpose.shape)\n",
    "\n",
    "#         print(\"\\ndense_kernel shape: \\n\",dense_kernel.shape,\"\\n\")\n",
    "\n",
    "        matmul_flatt_rel_dense_kernel = np.matmul(flatten_relu_out_transpose, dense_kernel)\n",
    "#         print(\"\\nmatmul_soft_dense_kernel shape: \\n\",matmul_flatt_rel_dense_kernel.shape,\"\\n\")\n",
    "#         print(\"\\nmatmul_soft_dense_kernel: \\n\",matmul_flatt_rel_dense_kernel,\"\\n\")\n",
    "\n",
    "        dense_bias_array = np.array(dense_bias)\n",
    "        dense_bias_array = dense_bias_array.reshape(1, NUM_CLASSES)  # 9 for 9 class\n",
    "#         print(\"\\ndense_bias_array: \\n\",dense_bias_array,\"\\n\")\n",
    "\n",
    "        add_matmul_flatt_rel_dense_kernel_and_dense_bias_array = matmul_flatt_rel_dense_kernel + dense_bias_array\n",
    "\n",
    "#         print(\"\\nvalue add_matmul_flatt_rel_dense_kernel_and_dense2_array: \\n\",add_matmul_flatt_rel_dense_kernel_and_dense_bias_array)\n",
    "\n",
    "        def softmax_fn(input_array):\n",
    "            e_x = np.exp(input_array - np.max(input_array))\n",
    "            return e_x / e_x.sum(axis=len(e_x.shape) - 1)\n",
    "\n",
    "        op = softmax_fn(add_matmul_flatt_rel_dense_kernel_and_dense_bias_array)\n",
    "        print(\"output of FC layer: \",op,\"\\n\")\n",
    "\n",
    "        # =============================================================================\n",
    "        #     Following code for finding class\n",
    "        # =============================================================================\n",
    "\n",
    "        m = 0\n",
    "        k = 0\n",
    "        # op=[[0.17095664, 0.24349895, 0.172376,   0.19243606, 0.62073235]]\n",
    "        # op=np.array(op)\n",
    "        # print(op.shape)\n",
    "        # print(type(op))\n",
    "\n",
    "        for h in op:\n",
    "\n",
    "            for index, j in enumerate(h):\n",
    "\n",
    "                o = j\n",
    "                # print(o)\n",
    "                if o > m:\n",
    "                    m = o\n",
    "                    #                print(m)\n",
    "                    k = index\n",
    "                else:\n",
    "                    pass\n",
    "        print('class:', k)\n",
    "        end = time.time()\n",
    "        elapsed_time = round((end - start) * 1000, 3)\n",
    "        total_time += elapsed_time\n",
    "        # normally gives time in second. multiply or divide to change unit of time\n",
    "        print(\"\\nElapsed Time: \", elapsed_time, \" milliseconds and total time is: \", total_time, \" milliseconds\")\n",
    "        print(\"\\n-------------------------------------------------------------------------------------------------\")\n",
    "\n",
    "    return image_number, total_time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Name of loaded image:  /media/atif/BE0E05910E0543BD/University of Bremen MSc/item_project/Image-classification/test_image/20_speed.ppm\n",
      "img shape:  (48, 48)  len fo shape:  2  ndim:  2\n",
      "conv kernel shape:  (5, 3, 3)  len fo shape:  3  ndim:  3\n",
      "in func:  (48, 48) \t (5, 3, 3)\n",
      "feature_shape:  (46, 46, 5)  feature ndim:  3\n",
      "\n",
      "relu_out_transpose shape:  (5, 46, 46)\n",
      "output of FC layer:  [[2.01219049e-04 6.05785105e-01 1.56110293e-04 1.47713526e-08\n",
      "  6.11215573e-03 2.60555872e-05 3.64587386e-01 5.76833987e-03\n",
      "  1.73636137e-02]] \n",
      "\n",
      "class: 1\n",
      "\n",
      "Elapsed Time:  1020.799  milliseconds and total time is:  185142.094  milliseconds\n",
      "\n",
      "-------------------------------------------------------------------------------------------------\n",
      "\n",
      "Name of loaded image:  /media/atif/BE0E05910E0543BD/University of Bremen MSc/item_project/Image-classification/test_image/30_speed.ppm\n",
      "img shape:  (48, 48)  len fo shape:  2  ndim:  2\n",
      "conv kernel shape:  (5, 3, 3)  len fo shape:  3  ndim:  3\n",
      "in func:  (48, 48) \t (5, 3, 3)\n",
      "feature_shape:  (46, 46, 5)  feature ndim:  3\n",
      "\n",
      "relu_out_transpose shape:  (5, 46, 46)\n",
      "output of FC layer:  [[2.30943039e-03 6.11381905e-01 2.95771516e-01 6.20261214e-06\n",
      "  3.30262088e-02 5.70476743e-02 3.00454404e-04 4.41052645e-06\n",
      "  1.52197961e-04]] \n",
      "\n",
      "class: 1\n",
      "\n",
      "Elapsed Time:  998.924  milliseconds and total time is:  186141.018  milliseconds\n",
      "\n",
      "-------------------------------------------------------------------------------------------------\n",
      "\n",
      "Total image number is:  2\n",
      "\n",
      "Total time for all classification:  186141.018  milliseconds\n",
      "\n",
      "Average time taken for per image classification:  93070.509  milliseconds\n"
     ]
    }
   ],
   "source": [
    "image_number, total_time = do_math(img_path, convolution_kernel_filter, conv_bias_array, IMG_SIZE,number_filter, NUM_CLASSES, total_time,dense_kernel, dense_bias)\n",
    "    \n",
    "print(\"\\nTotal image number is: \",image_number+1)\n",
    "print(\"\\nTotal time for all classification: \",total_time,\" milliseconds\")\n",
    "total_image=image_number+1\n",
    "print(\"\\nAverage time taken for per image classification: \",round((total_time/total_image),3),\" milliseconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [],
   "source": [
    "pic = np.zeros((5,5))\n",
    "fil = np.ones((3,3))\n",
    "pic = pic+2\n",
    "result = np.zeros((pic.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "filter_size= fil.shape[1]\n",
    "img_1 = pic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n",
      "18.0\n"
     ]
    }
   ],
   "source": [
    "for x in range(2):  # to get the value of convolution bias\n",
    "#         print(\"i am the conv_bias: \",x)\n",
    "        \n",
    "        for r in np.uint16(np.arange(filter_size/2.0,img_1.shape[0]-filter_size/2.0+1)):\n",
    "        \n",
    "            for c in np.uint16(np.arange(filter_size/2.0,img_1.shape[1]-filter_size/2.0+1)):\n",
    "            \n",
    "                curr_region = img_1[r-np.uint16(np.floor(filter_size/2.0)):r+np.uint16(np.ceil(filter_size/2.0)), c-np.uint16(np.floor(filter_size/2.0)):c+np.uint16(np.ceil(filter_size/2.0))]\n",
    "        \n",
    "                #Element-wise multipliplication between the current region and the filter.\n",
    "            \n",
    "                curr_result = curr_region * fil\n",
    "                conv_sum = np.sum(curr_result) #Summing the result of multiplication.\n",
    "                result[r, c] = conv_sum#Saving the summation in the convolution layer feature map.\n",
    "                print(result[r,c])\n",
    "        final_result = result[np.uint16(filter_size/2.0):result.shape[0]-np.uint16(filter_size/2.0),np.uint16(filter_size/2.0):result.shape[1]-np.uint16(filter_size/2.0)]\n",
    "#    print(\"\\nconv_ function finish\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[18., 18., 18.],\n",
       "       [18., 18., 18.],\n",
       "       [18., 18., 18.]])"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Extract weight from trained model file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from keras.models import load_model\n",
    "\n",
    "model = load_model('/home/atif/image_classification_c++/multi_filter_cpp/traffic_2_filter_no_pad_gray_ep_100_for_cpp.h5')\n",
    "#model = load_model('/home/atif/traffic_model_11_dec_1_filter.h5')\n",
    "layer_list =[]\n",
    "# f = open('/home/atif/path_for_storing_all_layer_info.txt', 'w') #uncomment it if you want to store all layer info at a time.\n",
    "for layer in model.layers:\n",
    "    g=layer.get_config()\n",
    "    h=layer.get_weights()\n",
    "    \n",
    "    layer_list.append(h)\n",
    "#     print (\"g== \",g,\"\\n\") #for printing layer name and verbal info\n",
    "\n",
    "#     print (\"h== \",h,\"\\n\\n\") # for printing layer numeric value, eg: weight, bias value\n",
    "#     print(\"type_of g == \",type(g),\"\\n\")\n",
    "#     print(\"type_of h == \",type(h),\"\\n\")\n",
    "\n",
    "# below lines till f.close() used for writing in text file. To do this you have to uncomment the above line started with f.open() also.\n",
    "\n",
    "#     g1=str(g) # declaring a string variable g1 to store the info of g\n",
    "#     h1=str(h) #declaring a string variable h1 to store the info of h\n",
    "#     g_type=str(type(g)) #declaring a string variable g1 to store the type of g\n",
    "#     h_type=str(type(h)) #declaring a string variable h1 to store the type of h\n",
    "    \n",
    "#     f.write(\"layer_definition: \"+g1+\"\\n\\n\")\n",
    "#     f.write(\"layer_type: \"+g_type+\"\\n\\n\")\n",
    "#     #f.write(\"\\n\")\n",
    "#     f.write(\"layer_weight: \"+h1+\"\\n\\n\")\n",
    "#     f.write(\"weight_type: \"+h_type+\"\\n\\n\\n\")\n",
    "#     f.write(\"\\n\")\n",
    "    \n",
    "# f.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Display extracted weight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "conv_kernel=layer_list[0][0]\n",
    "conv_kernel=conv_kernel.transpose()\n",
    "#print(\"conv_kernel: \\n\",conv_kernel,\"\\n\\n\")\n",
    "#print(\"conv_kernel shape:\\t\",conv_kernel.shape,\"\\n\\n\")\n",
    "#print(\"conv kernel dimension:\\t\",conv_kernel.ndim,\"\\n\\n\")\n",
    "#print(\"type_conv_kernel:\",type(conv_kernel),\"\\n\")\n",
    "\n",
    "\n",
    "\n",
    "conv_bias=layer_list[0][1]\n",
    "#print(\"conv_bias_value: \",conv_bias)\n",
    "#print(\"conv_bias ndim: \",conv_bias.ndim,\"\\n\\n\")\n",
    "\n",
    "\n",
    "\n",
    "dense_kernel=layer_list[2][0]\n",
    "#print(\"dense_kernel: \\n\",dense_kernel,\"\\n\\n\")\n",
    "#print(\"dense_kernel shape:\\t\",dense_kernel.shape,\"\\n\\n\")\n",
    "#print(\"dense_kernel dimension:\\t\",dense_kernel.ndim,\"\\n\\n\")\n",
    "#print(\"type_dense_kernel:\",type(dense_kernel),\"\\n\")\n",
    "#print(\"dense_kernel size: \",dense_kernel.size,\"\\n\")\n",
    "# dense_1_transpose=dense__1.transpose()\n",
    "# print(\"dense_1_transpose: \",dense_1_transpose,\"\\n\\n\")\n",
    "\n",
    "\n",
    "dense_bias=layer_list[2][1]\n",
    "#print(\"dense_bias: \",dense_bias)\n",
    "#print(\"dense_bias_shape: \",dense_bias.shape)\n",
    "dense_bias=dense_bias.reshape(1,9) # here chenge 5 to the number of your used class\n",
    "#print(\"dense_bias_shape: \",dense_bias.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Storing convolution kernel weight in a text file to use in CPP code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "conv_kernel=layer_list[0][0]\n",
    "conv_kernel=conv_kernel.transpose() # This has made to print it like a Matrix form. \n",
    "i_list=[]\n",
    "for i in conv_kernel:\n",
    "#     print(i)\n",
    "    i_list.append(i)\n",
    "#     for k in i:\n",
    "#         print(k)\n",
    "#         i_list.append(k)\n",
    "# print(i_list)\n",
    "i_list_array=[]\n",
    "i_list_array=np.array(i_list)\n",
    "# print(i_list_array.shape)\n",
    "# i_list_array=i_list_array.reshape(2,3,3)\n",
    "# print(i_list_array.ndim)\n",
    "\n",
    "for p in i_list_array:\n",
    "#     for a in p:\n",
    "#         print(a)\n",
    "#    print(p)\n",
    "    ww=str(p)\n",
    "    ww=ww.replace('[','')\n",
    "    ww=ww.replace(']','')\n",
    "    #f=open('/home/atif/conv_k_spy.txt','a') #uncomment from here till f.close() if you want to save text file\n",
    "    #f.write(ww)\n",
    "    #f.write(\"\\n\")\n",
    "#f.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Storing convolution kernel bias value in a text file to use in CPP code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "conv_bias=layer_list[0][1]\n",
    "conv_bias_list=[]\n",
    "for i in conv_bias:\n",
    "    conv_bias_list.append(i)\n",
    "\n",
    "conv_bias_array=[]\n",
    "conv_bias_array=np.array(conv_bias_list)\n",
    "#print(conv_bias_array)\n",
    "#np.savetxt('/home/atif/conv_b_spy.txt', conv_bias_array, fmt='%1.8e',delimiter=' ')\n",
    "\n",
    "#for x in conv_bias_array:\n",
    "    #print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Storing Dense kernel weight in a text file to use in CPP code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "dense_kernel=layer_list[2][0]\n",
    "i_list=[] #declare a list to store the weight of dense kernel\n",
    "for i in dense_kernel:\n",
    "#     print(i)\n",
    "    i_list.append(i) #appended it in the declared list\n",
    "#     for k in i:\n",
    "# #         print(k)\n",
    "#         i_list.append(k)\n",
    "# print(i_list)\n",
    "i_list_array=[] #declared an array\n",
    "i_list_array=np.array(i_list) # store the value of list in the array\n",
    "#print(i_list_array)\n",
    "#np.savetxt('/home/atif/dense_k_spy.txt', i_list_array, fmt='%1.8e',delimiter=' ') #writing on a text file from array\n",
    "\n",
    "# %.8f #you can use it to get float value\n",
    "# fmt='%1.8e' #add this above line after i_list_aray"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Storing Dense kernel bias value in a text file to use in CPP code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "dense_bias=layer_list[2][1]\n",
    "dense_bias_list=[]\n",
    "for i in dense_bias:\n",
    "    dense_bias_list.append(i)\n",
    "\n",
    "dense_bias_array=[]\n",
    "dense_bias_array=np.array(dense_bias_list)\n",
    "#print(dense_bias_array)\n",
    "#np.savetxt('/home/atif/dense_b_spy.txt', dense_bias_array, fmt='%1.8e',delimiter=' ')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reshaping convolution kernel for further process"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#print(\"conv_kernel:\\n\",conv_kernel,\"\\n\")\n",
    "#print(\"conv_kernel_shape:\",conv_kernel.shape,\"\\tconv_kernel ndim:\",conv_kernel.ndim,\"\\n\")\n",
    "#print(\"length of conv_kernel:\",len(conv_kernel),\"\\n\")\n",
    "\n",
    "conv_kernel_reshape=conv_kernel.reshape(2,3,3) # 2 for 2 filter. change it according to your filter number\n",
    "#print(\"conv_kernel_reshape:\\n\",conv_kernel_reshape,\"\\n\")\n",
    "#print(\"conv_kernel_reshape shape:\",conv_kernel_reshape.shape,\"\\tconv_kernel_reshape ndim:\",conv_kernel_reshape.ndim,\"\\n\")\n",
    "#print(\"length of conv_kernel_reshape:\",len(conv_kernel_reshape[0]),\"\\n\")\n",
    "\n",
    "convolution_kernel_filter=[]\n",
    "convolution_kernel_filter=np.zeros((2,3,3)) # 2 for 2 filter. change it according to your filter number\n",
    "convolution_kernel_filter[:,:,:]=np.array(conv_kernel_reshape)\n",
    "#print(\"convolution_kernel_filter: \\n\",convolution_kernel_filter,\"\\n\")\n",
    "#print(\"convolution_kernel_filter shape:\",convolution_kernel_filter.shape,\"\\tconvolution_kernel_filter ndim:\",convolution_kernel_filter.ndim,\"\\n\")\n",
    "#print(\"length of convolution_kernel_filter:\",len(convolution_kernel_filter),\"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Steps for Classification\n",
    "## At first processing image, Convolution and apply Relu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Name of loaded image:  /home/atif/image_classification_c++/multi_filter_cpp/test_image/50_speed.ppm\n",
      "Filter  1\n",
      "Filter  2\n",
      "3.466926483846102e-142\n",
      "2.3085899666969084e-48\n",
      "6.995688263318184e-30\n",
      "3.8409075116573936e-15\n",
      "0.9999999999999962\n",
      "class: 4\n",
      "\n",
      "Name of loaded image:  /home/atif/image_classification_c++/multi_filter_cpp/test_image/20_speed.ppm\n",
      "Filter  1\n",
      "Filter  2\n",
      "1.4442679059689178e-133\n",
      "2.518911831904704e-30\n",
      "1.3016192821246158e-15\n",
      "0.046888826826912006\n",
      "0.9531111731730867\n",
      "class: 4\n",
      "\n",
      "Name of loaded image:  /home/atif/image_classification_c++/multi_filter_cpp/test_image/stop.ppm\n",
      "Filter  1\n",
      "Filter  2\n",
      "3.380052453539173e-138\n",
      "9.902191131949987e-56\n",
      "1.4209070255903226e-31\n",
      "1.0\n",
      "class: 3\n"
     ]
    }
   ],
   "source": [
    "def preprocess_img(img):\n",
    "#     uncomment following 5 lines for rgb testing and comment out the rgb2gray line\n",
    "#     Histogram normalization in y\n",
    "#     hsv = color.rgb2hsv(img)\n",
    "#     hsv[:,:,2] = exposure.equalize_hist(hsv[:,:,2])\n",
    "#     img = color.hsv2rgb(hsv)\n",
    "\n",
    "    # central scrop\n",
    "    min_side = min(img.shape[:-1])\n",
    "    centre = img.shape[0]//2, img.shape[1]//2\n",
    "    img = img[centre[0]-min_side//2:centre[0]+min_side//2,\n",
    "              centre[1]-min_side//2:centre[1]+min_side//2,\n",
    "              :]\n",
    "    img = rgb2gray(img) #rgb to gray conversion\n",
    "\n",
    "    # rescale to standard size\n",
    "    img = transform.resize(img, (IMG_SIZE, IMG_SIZE))\n",
    "\n",
    "    # roll color axis to axis 0\n",
    "    img = np.rollaxis(img,-1)\n",
    "\n",
    "    return img\n",
    "\n",
    "def conv_(img, conv_filter):\n",
    "#     print(\"\\nconv_ function start to work\\n\")\n",
    "    filter_size = conv_filter.shape[1] #output is 3\n",
    "#     print(\"filter_size: \",filter_size)\n",
    "#     print(\"img shape: \",img.shape)\n",
    "    result = np.zeros((img.shape))\n",
    "    #Looping through the image to apply the convolution operation.\n",
    "    \n",
    "    for x in conv_bias_array:  # to get the value of convolution bias\n",
    "#         print(\"i am the conv_bias: \",x)\n",
    "        \n",
    "        for r in np.uint16(np.arange(filter_size/2.0,img.shape[0]-filter_size/2.0+1)):\n",
    "        \n",
    "            for c in np.uint16(np.arange(filter_size/2.0,img.shape[1]-filter_size/2.0+1)):\n",
    "            \n",
    "                curr_region = img[r-np.uint16(np.floor(filter_size/2.0)):r+np.uint16(np.ceil(filter_size/2.0)), c-np.uint16(np.floor(filter_size/2.0)):c+np.uint16(np.ceil(filter_size/2.0))]\n",
    "        \n",
    "                #Element-wise multipliplication between the current region and the filter.\n",
    "            \n",
    "                curr_result = curr_region * conv_filter\n",
    "#             print(\"curr_result: \",curr_result)\n",
    "                curr_result= curr_result+x\n",
    "                \n",
    "#             print(\"conv_bias: \",conv_bias_new)\n",
    "#                 print(\"new curr res: \",curr_result)\n",
    "                conv_sum = np.sum(curr_result) #Summing the result of multiplication.\n",
    "                result[r, c] = conv_sum#Saving the summation in the convolution layer feature map.\n",
    "#             print(\"conv_sum_shape: \",conv_sum.shape)\n",
    "#         print(\"now x is: \",x)\n",
    "        #print(curr_region)\n",
    "    #Clipping the outliers of the result matrix.\n",
    "    final_result = result[np.uint16(filter_size/2.0):result.shape[0]-np.uint16(filter_size/2.0),np.uint16(filter_size/2.0):result.shape[1]-np.uint16(filter_size/2.0)]\n",
    "#     print(\"\\nconv_ function finish\\n\")\n",
    "    return final_result\n",
    "\n",
    "\n",
    "\n",
    "def conv(img, conv_filter):\n",
    "#     print(\"\\nconv function start to work\")\n",
    "    \n",
    "    if len(img.shape) > 2 or len(conv_filter.shape) > 3: # Check if number of image channels matches the filter depth.\n",
    "        if img.shape[-1] != conv_filter.shape[-1]:\n",
    "            print(\"Error: Number of channels in both image and filter must match.\")\n",
    "            sys.exit()\n",
    "    if conv_filter.shape[1] != conv_filter.shape[2]: # Check if filter dimensions are equal.\n",
    "        print('Error: Filter must be a square matrix. I.e. number of rows and columns must match.')\n",
    "        sys.exit()\n",
    "    if conv_filter.shape[1]%2==0: # Check if filter diemnsions are odd.\n",
    "        print('Error: Filter must have an odd size. I.e. number of rows and columns must be odd.')\n",
    "        sys.exit()\n",
    "\n",
    "    # An empty feature map to hold the output of convolving the filter(s) with the image.\n",
    "    feature_maps = np.zeros((img.shape[0]-conv_filter.shape[1]+1, img.shape[1]-conv_filter.shape[1]+1, conv_filter.shape[0]))\n",
    "\n",
    "    # Convolving the image by the filter(s).\n",
    "    \n",
    "    for filter_num in range(conv_filter.shape[0]):\n",
    "#         print(\"filter num: \",filter_num)\n",
    "        print(\"Filter \", filter_num + 1)\n",
    "        curr_filter = conv_filter[filter_num, :] # getting a filter from the bank.\n",
    "#         print(curr_filter)\n",
    "        #print(\"curr_fiter_shape: \",curr_filter.shape)\n",
    "#         print(\"length of curr_fiter_shape: \",len(curr_filter.shape))\n",
    "\n",
    "\n",
    "        if len(curr_filter.shape) > 2:\n",
    "            conv_map = conv_(img[:, :, 0], curr_filter[:, :, 0]) # Array holding the sum of all feature maps.\n",
    "            for ch_num in range(1, curr_filter.shape[-1]): # Convolving each channel with the image and summing the results.\n",
    "                conv_map = conv_map + conv_(img[:, :, ch_num], \n",
    "                                  curr_filter[:, :, ch_num])\n",
    "        else: # There is just a single channel in the filter.\n",
    "#             print(\"\\nGo to conv_ function \")\n",
    "            conv_map = conv_(img, curr_filter)\n",
    "            \n",
    "        feature_maps[:, :, filter_num] = conv_map # Holding feature map with the current filter.\n",
    "#         print(\"feature_maps from conv_map: \",feature_maps)\n",
    "    return feature_maps # Returning all feature maps.\n",
    "\n",
    "\n",
    "def relu(feature_map):\n",
    "    #Preparing the output of the ReLU activation function.\n",
    "    relu_out = np.zeros(feature_map.shape)\n",
    "    for map_num in range(feature_map.shape[-1]):\n",
    "        for r in np.arange(0,feature_map.shape[0]):\n",
    "            for c in np.arange(0, feature_map.shape[1]):\n",
    "                relu_out[r, c, map_num] = np.max([feature_map[r, c, map_num], 0])\n",
    "    return relu_out\n",
    "\n",
    "\n",
    "\n",
    "path = r'/home/atif/image_classification_c++/multi_filter_cpp/test_image/'\n",
    "\n",
    "img_path = glob.glob(path+ '/*.ppm')\n",
    "for image in img_path:\n",
    "    print(\"\\nName of loaded image: \",image)\n",
    "    X_test=[]\n",
    "    X_test.append(preprocess_img(io.imread(image)))\n",
    "    X_test = np.array(X_test)\n",
    "#     plt.imshow(X_test)\n",
    "    X_test = X_test.reshape(IMG_SIZE,IMG_SIZE)\n",
    "    \n",
    "    feature=conv(img=X_test,conv_filter=convolution_kernel_filter) #conv function calling\n",
    "    relu_out=relu(feature) # relu function calling\n",
    "    \n",
    "    # output of feature map / conv function\n",
    "\n",
    "#     print(\"\\nfeature shape: \\n\",feature.shape)\n",
    "\n",
    "    # x_feature_map=np.flipud(feature[0])\n",
    "    transpose_feature_map=feature.transpose()\n",
    "#     print(\"\\ntranspose_feature_map shape: \",transpose_feature_map.shape)\n",
    "#     plt.imshow(transpose_feature_map[0])\n",
    "#     print(\"\\ntranspose_feature_map: \\n\",transpose_feature_map)\n",
    "    \n",
    "#     print(\"\\nrelu_out shape: \",relu_out.shape)\n",
    "    relu_out_transpose=relu_out.transpose()\n",
    "#     print(\"\\nrelu_out_transpose shape: \",relu_out_transpose.shape)\n",
    "#     plt.imshow(relu_out_transpose[0])\n",
    "#     print(\"\\nrelu_out_transpose:\\n\",relu_out_transpose)\n",
    "    \n",
    "# Matrix Multiplication with Dense kernel weight and Convolved Image\n",
    "\n",
    "    flatten_relu_out_transpose=relu_out_transpose.reshape(1,2*46*46)  #if you don't do padd on input image please make it 46*46. how 46 came? \n",
    "                                                                                    #the formula of output size. and 2 for 2 feature map. Here 2 filter so 2 featue map\n",
    "#     print(\"\\nflatten_relu_out_transpose shape: \\n\",flatten_relu_out_transpose.shape)\n",
    "#     print(\"\\ndense_kernel shape: \\n\",dense_kernel.shape,\"\\n\")\n",
    "    matmul_flatt_rel_dense_kernel=np.matmul(flatten_relu_out_transpose,dense_kernel)\n",
    "#     print(\"\\nmatmul_soft_dense_kernel shape: \\n\",matmul_flatt_rel_dense_kernel.shape,\"\\n\")\n",
    "#     print(\"\\nmatmul_soft_dense_kernel: \\n\",matmul_flatt_rel_dense_kernel,\"\\n\")\n",
    "\n",
    "    dense_bias_array=np.array(dense_bias)\n",
    "    dense_bias_array=dense_bias_array.reshape(1,9) # 9 for 9 class\n",
    "#     print(\"\\ndense_bias_array: \\n\",dense_bias_array,\"\\n\")\n",
    "\n",
    "    add_matmul_flatt_rel_dense_kernel_and_dense_bias_array=matmul_flatt_rel_dense_kernel+dense_bias_array\n",
    "#     print(\"\\nvalue add_matmul_flatt_rel_dense_kernel_and_dense2_array: \\n\",add_matmul_flatt_rel_dense_kernel_and_dense_bias_array)\n",
    "    \n",
    "    def softmax_fn(input_array):\n",
    "        e_x=np.exp(input_array-np.max(input_array))\n",
    "        return e_x/e_x.sum(axis=len(e_x.shape)-1)\n",
    "\n",
    "    op= softmax_fn(add_matmul_flatt_rel_dense_kernel_and_dense_bias_array)\n",
    "#     print(\"output of FC layer: \",op,\"\\n\")\n",
    "    \n",
    "#   Show class of loaded image\n",
    "\n",
    "    m=0\n",
    "    k=0\n",
    "    # op=[[0.17095664, 0.24349895, 0.172376,   0.19243606, 0.62073235]]\n",
    "    # op=np.array(op)\n",
    "    # print(op.shape)\n",
    "    # print(type(op))\n",
    "\n",
    "    for h in op:\n",
    "        for index,j in enumerate(h):\n",
    "            o=j\n",
    "            #print(o)\n",
    "            if o>m:\n",
    "                m=o\n",
    "                print(m)\n",
    "                k=index\n",
    "            else:\n",
    "                pass\n",
    "    print('class:',k)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}