Step2_DIP.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "C:\\Users\\owner\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import Input, Dense, Permute, multiply, GlobalAveragePooling1D, Reshape, concatenate\n",
    "import tensorflow.keras.backend as K\n",
    "import numpy as np\n",
    "from keras import optimizers, Model\n",
    "from keras.utils import Sequence\n",
    "from numpy.random import seed\n",
    "import math\n",
    "from keras.callbacks import EarlyStopping,ModelCheckpoint\n",
    "\n",
    "\n",
    "def squeeze_excite_block1(tensor, ratio=16):\n",
    "    init = tensor\n",
    "    channel_axis = 1 if K.image_data_format() == \"channels_first\" else -1\n",
    "    filters = init._keras_shape[channel_axis]\n",
    "    se_shape = (1,filters)\n",
    "    init1 = Reshape(target_shape=(se_shape))(init)    \n",
    "    se = GlobalAveragePooling1D()(init1)\n",
    "    se = Reshape(se_shape)(se)\n",
    "    se = Dense(filters // ratio, activation='relu', kernel_initializer='he_normal', use_bias=False)(se)\n",
    "    se = Dense(filters, activation='sigmoid', kernel_initializer='he_normal', use_bias=False)(se)\n",
    "\n",
    "    \n",
    "    if K.image_data_format() == 'channels_first':\n",
    "        se = Permute((3, 1, 2))(se) \n",
    "    se_reshape = (filters,)\n",
    "    se = Reshape(target_shape=(se_reshape))(se)\n",
    "    x = multiply([init, se])\n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_4\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_4 (InputLayer)            (None, 3375)         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "e1 (Dense)                      (None, 128)          432128      input_4[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "e2 (Dense)                      (None, 64)           8256        e1[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "e3 (Dense)                      (None, 32)           2080        e2[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "e4 (Dense)                      (None, 16)           528         e3[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "e5 (Dense)                      (None, 8)            136         e4[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "e6 (Dense)                      (None, 4)            36          e5[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "d1 (Dense)                      (None, 4)            20          e6[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_127 (Reshape)           (None, 1, 4)         0           d1[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_124 (Reshape)           (None, 1, 4)         0           e6[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_43 (Gl (None, 4)            0           reshape_127[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_42 (Gl (None, 4)            0           reshape_124[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_128 (Reshape)           (None, 1, 4)         0           global_average_pooling1d_43[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "reshape_125 (Reshape)           (None, 1, 4)         0           global_average_pooling1d_42[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "dense_88 (Dense)                (None, 1, 1)         4           reshape_128[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_86 (Dense)                (None, 1, 1)         4           reshape_125[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_89 (Dense)                (None, 1, 4)         4           dense_88[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_87 (Dense)                (None, 1, 4)         4           dense_86[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_129 (Reshape)           (None, 4)            0           dense_89[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_126 (Reshape)           (None, 4)            0           dense_87[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "multiply_43 (Multiply)          (None, 4)            0           d1[0][0]                         \n",
      "                                                                 reshape_129[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "multiply_42 (Multiply)          (None, 4)            0           e6[0][0]                         \n",
      "                                                                 reshape_126[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_19 (Concatenate)    (None, 8)            0           multiply_43[0][0]                \n",
      "                                                                 multiply_42[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "d2 (Dense)                      (None, 8)            72          concatenate_19[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "reshape_130 (Reshape)           (None, 1, 8)         0           d2[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_121 (Reshape)           (None, 1, 8)         0           e5[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_44 (Gl (None, 8)            0           reshape_130[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_41 (Gl (None, 8)            0           reshape_121[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_131 (Reshape)           (None, 1, 8)         0           global_average_pooling1d_44[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "reshape_122 (Reshape)           (None, 1, 8)         0           global_average_pooling1d_41[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "dense_90 (Dense)                (None, 1, 1)         8           reshape_131[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_84 (Dense)                (None, 1, 1)         8           reshape_122[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_91 (Dense)                (None, 1, 8)         8           dense_90[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_85 (Dense)                (None, 1, 8)         8           dense_84[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_132 (Reshape)           (None, 8)            0           dense_91[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_123 (Reshape)           (None, 8)            0           dense_85[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "multiply_44 (Multiply)          (None, 8)            0           d2[0][0]                         \n",
      "                                                                 reshape_132[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "multiply_41 (Multiply)          (None, 8)            0           e5[0][0]                         \n",
      "                                                                 reshape_123[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_20 (Concatenate)    (None, 16)           0           multiply_44[0][0]                \n",
      "                                                                 multiply_41[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "d3 (Dense)                      (None, 16)           272         concatenate_20[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "reshape_133 (Reshape)           (None, 1, 16)        0           d3[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_118 (Reshape)           (None, 1, 16)        0           e4[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_45 (Gl (None, 16)           0           reshape_133[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_40 (Gl (None, 16)           0           reshape_118[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_134 (Reshape)           (None, 1, 16)        0           global_average_pooling1d_45[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "reshape_119 (Reshape)           (None, 1, 16)        0           global_average_pooling1d_40[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "dense_92 (Dense)                (None, 1, 1)         16          reshape_134[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_82 (Dense)                (None, 1, 1)         16          reshape_119[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_93 (Dense)                (None, 1, 16)        16          dense_92[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_83 (Dense)                (None, 1, 16)        16          dense_82[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_135 (Reshape)           (None, 16)           0           dense_93[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_120 (Reshape)           (None, 16)           0           dense_83[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "multiply_45 (Multiply)          (None, 16)           0           d3[0][0]                         \n",
      "                                                                 reshape_135[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "multiply_40 (Multiply)          (None, 16)           0           e4[0][0]                         \n",
      "                                                                 reshape_120[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_21 (Concatenate)    (None, 32)           0           multiply_45[0][0]                \n",
      "                                                                 multiply_40[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "d4 (Dense)                      (None, 32)           1056        concatenate_21[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "reshape_136 (Reshape)           (None, 1, 32)        0           d4[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_115 (Reshape)           (None, 1, 32)        0           e3[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_46 (Gl (None, 32)           0           reshape_136[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_39 (Gl (None, 32)           0           reshape_115[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_137 (Reshape)           (None, 1, 32)        0           global_average_pooling1d_46[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "reshape_116 (Reshape)           (None, 1, 32)        0           global_average_pooling1d_39[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "dense_94 (Dense)                (None, 1, 1)         32          reshape_137[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_80 (Dense)                (None, 1, 1)         32          reshape_116[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_95 (Dense)                (None, 1, 32)        32          dense_94[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_81 (Dense)                (None, 1, 32)        32          dense_80[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_138 (Reshape)           (None, 32)           0           dense_95[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_117 (Reshape)           (None, 32)           0           dense_81[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "multiply_46 (Multiply)          (None, 32)           0           d4[0][0]                         \n",
      "                                                                 reshape_138[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "multiply_39 (Multiply)          (None, 32)           0           e3[0][0]                         \n",
      "                                                                 reshape_117[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_22 (Concatenate)    (None, 64)           0           multiply_46[0][0]                \n",
      "                                                                 multiply_39[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "d5 (Dense)                      (None, 64)           4160        concatenate_22[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "reshape_139 (Reshape)           (None, 1, 64)        0           d5[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_112 (Reshape)           (None, 1, 64)        0           e2[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_47 (Gl (None, 64)           0           reshape_139[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_38 (Gl (None, 64)           0           reshape_112[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_140 (Reshape)           (None, 1, 64)        0           global_average_pooling1d_47[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "reshape_113 (Reshape)           (None, 1, 64)        0           global_average_pooling1d_38[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "dense_96 (Dense)                (None, 1, 1)         64          reshape_140[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_78 (Dense)                (None, 1, 1)         64          reshape_113[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_97 (Dense)                (None, 1, 64)        64          dense_96[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_79 (Dense)                (None, 1, 64)        64          dense_78[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_141 (Reshape)           (None, 64)           0           dense_97[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_114 (Reshape)           (None, 64)           0           dense_79[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "multiply_47 (Multiply)          (None, 64)           0           d5[0][0]                         \n",
      "                                                                 reshape_141[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "multiply_38 (Multiply)          (None, 64)           0           e2[0][0]                         \n",
      "                                                                 reshape_114[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_23 (Concatenate)    (None, 128)          0           multiply_47[0][0]                \n",
      "                                                                 multiply_38[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "d6 (Dense)                      (None, 128)          16512       concatenate_23[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "reshape_142 (Reshape)           (None, 1, 128)       0           d6[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "reshape_109 (Reshape)           (None, 1, 128)       0           e1[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_48 (Gl (None, 128)          0           reshape_142[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_37 (Gl (None, 128)          0           reshape_109[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "reshape_143 (Reshape)           (None, 1, 128)       0           global_average_pooling1d_48[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "reshape_110 (Reshape)           (None, 1, 128)       0           global_average_pooling1d_37[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "dense_98 (Dense)                (None, 1, 1)         128         reshape_143[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_76 (Dense)                (None, 1, 1)         128         reshape_110[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_99 (Dense)                (None, 1, 128)       128         dense_98[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_77 (Dense)                (None, 1, 128)       128         dense_76[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_144 (Reshape)           (None, 128)          0           dense_99[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_111 (Reshape)           (None, 128)          0           dense_77[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "multiply_48 (Multiply)          (None, 128)          0           d6[0][0]                         \n",
      "                                                                 reshape_144[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "multiply_37 (Multiply)          (None, 128)          0           e1[0][0]                         \n",
      "                                                                 reshape_111[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_24 (Concatenate)    (None, 256)          0           multiply_48[0][0]                \n",
      "                                                                 multiply_37[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_100 (Dense)               (None, 3375)         867375      concatenate_24[0][0]             \n",
      "==================================================================================================\n",
      "Total params: 1,333,639\n",
      "Trainable params: 1,333,639\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "36288/36288 [==============================] - 12s 330us/step - loss: 0.0592\n",
      "Epoch 2/50\n",
      "36288/36288 [==============================] - 10s 268us/step - loss: 0.0499\n",
      "Epoch 3/50\n",
      "36288/36288 [==============================] - 9s 251us/step - loss: 0.0467\n",
      "Epoch 4/50\n",
      "36288/36288 [==============================] - 10s 264us/step - loss: 0.0449\n",
      "Epoch 5/50\n",
      "36288/36288 [==============================] - 10s 267us/step - loss: 0.0438\n",
      "Epoch 6/50\n",
      "36288/36288 [==============================] - 9s 258us/step - loss: 0.0430\n",
      "Epoch 7/50\n",
      "36288/36288 [==============================] - 9s 261us/step - loss: 0.0424\n",
      "Epoch 8/50\n",
      "36288/36288 [==============================] - 9s 258us/step - loss: 0.0419\n",
      "Epoch 9/50\n",
      "36288/36288 [==============================] - 9s 258us/step - loss: 0.0416\n",
      "Epoch 10/50\n",
      "36288/36288 [==============================] - 9s 258us/step - loss: 0.0413\n",
      "Epoch 11/50\n",
      "36288/36288 [==============================] - 9s 259us/step - loss: 0.0410\n",
      "Epoch 12/50\n",
      "36288/36288 [==============================] - 9s 262us/step - loss: 0.0408\n",
      "Epoch 13/50\n",
      "36288/36288 [==============================] - 9s 261us/step - loss: 0.0406\n",
      "Epoch 14/50\n",
      "36288/36288 [==============================] - 9s 261us/step - loss: 0.0405\n",
      "Epoch 15/50\n",
      "36288/36288 [==============================] - 9s 260us/step - loss: 0.0404\n",
      "Epoch 16/50\n",
      "36288/36288 [==============================] - 9s 259us/step - loss: 0.0403\n",
      "Epoch 17/50\n",
      "36288/36288 [==============================] - 9s 258us/step - loss: 0.0402\n",
      "Epoch 18/50\n",
      "36288/36288 [==============================] - 10s 273us/step - loss: 0.0401\n",
      "Epoch 19/50\n",
      "36288/36288 [==============================] - 10s 269us/step - loss: 0.0400\n",
      "Epoch 20/50\n",
      "36288/36288 [==============================] - 9s 262us/step - loss: 0.0399\n",
      "Epoch 21/50\n",
      "36288/36288 [==============================] - 9s 260us/step - loss: 0.0399\n",
      "Epoch 22/50\n",
      "36288/36288 [==============================] - 10s 278us/step - loss: 0.0398\n",
      "Epoch 23/50\n",
      "36288/36288 [==============================] - 10s 274us/step - loss: 0.0398\n",
      "Epoch 24/50\n",
      "36288/36288 [==============================] - 10s 262us/step - loss: 0.0397\n",
      "Epoch 25/50\n",
      "36288/36288 [==============================] - 9s 261us/step - loss: 0.0397\n",
      "Epoch 26/50\n",
      "36288/36288 [==============================] - 9s 262us/step - loss: 0.0397\n",
      "Epoch 27/50\n",
      "36288/36288 [==============================] - 9s 262us/step - loss: 0.0396\n",
      "Epoch 28/50\n",
      "36288/36288 [==============================] - 10s 262us/step - loss: 0.0396\n",
      "Epoch 29/50\n",
      "36288/36288 [==============================] - 10s 262us/step - loss: 0.0396\n",
      "Epoch 30/50\n",
      "36288/36288 [==============================] - 10s 274us/step - loss: 0.0395\n",
      "Epoch 31/50\n",
      "36288/36288 [==============================] - 10s 267us/step - loss: 0.0395\n",
      "Epoch 32/50\n",
      "36288/36288 [==============================] - 10s 265us/step - loss: 0.0395\n",
      "Epoch 33/50\n",
      "36288/36288 [==============================] - 9s 261us/step - loss: 0.0395\n",
      "Epoch 34/50\n",
      "36288/36288 [==============================] - 9s 261us/step - loss: 0.0394\n",
      "Epoch 35/50\n",
      "36288/36288 [==============================] - 10s 265us/step - loss: 0.0394\n",
      "Epoch 36/50\n",
      "36288/36288 [==============================] - 10s 263us/step - loss: 0.0394\n",
      "Epoch 37/50\n",
      "36288/36288 [==============================] - 10s 263us/step - loss: 0.0394\n",
      "Epoch 38/50\n",
      "36288/36288 [==============================] - 10s 262us/step - loss: 0.0394\n",
      "Epoch 39/50\n",
      "36288/36288 [==============================] - 10s 263us/step - loss: 0.0393\n",
      "Epoch 40/50\n",
      "36288/36288 [==============================] - 10s 273us/step - loss: 0.0393\n",
      "Epoch 41/50\n",
      "36288/36288 [==============================] - 10s 264us/step - loss: 0.0393\n",
      "Epoch 42/50\n",
      "36288/36288 [==============================] - 10s 275us/step - loss: 0.0393\n",
      "Epoch 43/50\n",
      "36288/36288 [==============================] - 10s 267us/step - loss: 0.0393\n",
      "Epoch 44/50\n",
      "36288/36288 [==============================] - 10s 266us/step - loss: 0.0393\n",
      "Epoch 45/50\n",
      "36288/36288 [==============================] - 10s 264us/step - loss: 0.0393\n",
      "Epoch 46/50\n",
      "36288/36288 [==============================] - 10s 272us/step - loss: 0.0393\n",
      "Epoch 47/50\n",
      "36288/36288 [==============================] - 10s 277us/step - loss: 0.0393\n",
      "Epoch 48/50\n",
      "36288/36288 [==============================] - 10s 274us/step - loss: 0.0392\n",
      "Epoch 49/50\n",
      "36288/36288 [==============================] - 10s 277us/step - loss: 0.0392\n",
      "Epoch 50/50\n",
      "36288/36288 [==============================] - 10s 271us/step - loss: 0.0392\n"
     ]
    }
   ],
   "source": [
    "import h5py\n",
    "# Reading Patches.\n",
    "f = h5py.File('Syn_Example.mat', 'r')\n",
    "dataNoise= f.get('X')\n",
    "dataNoise = np.array(dataNoise)\n",
    "INPUT_SIZE1 = dataNoise.shape[0]\n",
    "INPUT_SIZE2 = dataNoise.shape[1]\n",
    "input_img = Input(shape=(INPUT_SIZE2,))\n",
    "\n",
    "# Building the DIP model.\n",
    "D1 = 128\n",
    "D2 = int(D1/2)\n",
    "D3 = int(D2/2)\n",
    "D4 = int(D3/2)\n",
    "D5 = int(D4/2)\n",
    "D6 = int(D5/2)\n",
    "D7 = int(D6/2)\n",
    "\n",
    "\n",
    "\n",
    "# Encoder\n",
    "encoded1 = Dense(D1,  activation='elu',name='e1')(input_img)\n",
    "encoded1x = squeeze_excite_block1(encoded1,D1)\n",
    "\n",
    "encoded2 = Dense(D2,  activation='elu',name='e2')(encoded1)\n",
    "encoded2x = squeeze_excite_block1(encoded2,D2)\n",
    "\n",
    "encoded3 = Dense(D3,  activation='elu',name='e3')(encoded2)\n",
    "encoded3x = squeeze_excite_block1(encoded3,D3)\n",
    "\n",
    "encoded4 = Dense(D4,  activation='elu',name='e4')(encoded3)\n",
    "encoded4x = squeeze_excite_block1(encoded4,D4)\n",
    "\n",
    "encoded5 = Dense(D5,  activation='elu',name='e5')(encoded4)\n",
    "encoded5x = squeeze_excite_block1(encoded5,D5)\n",
    "\n",
    "encoded6 = Dense(D6,  activation='elu',name='e6')(encoded5)\n",
    "encoded6x = squeeze_excite_block1(encoded6,D6)\n",
    "\n",
    "\n",
    "# Decoder\n",
    "decoded1aa = Dense(D6,  activation='elu',name='d1')(encoded6)\n",
    "decoded1aa = squeeze_excite_block1(decoded1aa,D6)\n",
    "decoded1aa = concatenate([decoded1aa,encoded6x])\n",
    "\n",
    "decoded1a = Dense(D5,  activation='elu',name='d2')(decoded1aa)\n",
    "decoded1a = squeeze_excite_block1(decoded1a,D5)\n",
    "decoded1a = concatenate([decoded1a,encoded5x])\n",
    "\n",
    "decoded1 = Dense(D4,  activation='elu',name='d3')(decoded1a)\n",
    "decoded1 = squeeze_excite_block1(decoded1,D4)\n",
    "decoded1 = concatenate([decoded1,encoded4x])\n",
    "\n",
    "decoded2 = Dense(D3,  activation='elu',name='d4')(decoded1)\n",
    "decoded2 = squeeze_excite_block1(decoded2,D3)\n",
    "decoded2 = concatenate([decoded2,encoded3x])\n",
    "\n",
    "decoded3 = Dense(D2,  activation='elu',name='d5')(decoded2)\n",
    "decoded3 = squeeze_excite_block1(decoded3,D2)\n",
    "decoded3 = concatenate([decoded3,encoded2x])\n",
    "\n",
    "decoded4 = Dense(D1,  activation='elu',name='d6')(decoded3)\n",
    "decoded4 = squeeze_excite_block1(decoded4,D1)\n",
    "decoded4 = concatenate([decoded4,encoded1x])\n",
    "\n",
    "# Final Layer\n",
    "decoded = Dense(INPUT_SIZE2, activation='linear')(decoded4)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "    \n",
    "\n",
    "            \n",
    "autoencoder = Model(input_img, decoded)\n",
    "sgd = optimizers.adam(lr=0.0001)\n",
    "autoencoder.compile(optimizer=sgd, loss='mse')\n",
    "\n",
    "autoencoder.summary()\n",
    "\n",
    "# Addig stopping condition\n",
    "es=EarlyStopping(monitor='loss',mode='min',verbose=1,patience=5)\n",
    "mc=ModelCheckpoint('best_model.h5',monitor='loss',mode='min',save_best_only=True)\n",
    "\n",
    "\n",
    "batch = 128\n",
    "history = autoencoder.fit(dataNoise,dataNoise, epochs=50, batch_size=batch,callbacks=[es,mc],shuffle=False)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Predict\n",
    "from keras.models import load_model\n",
    "import h5py\n",
    "import scipy.io\n",
    "\n",
    "model = load_model('best_model.h5')\n",
    "out = model.predict(dataNoise, batch_size=batch)\n",
    "scipy.io.savemat('Syn_Example_outpatch.mat', mdict={'out': out})\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}