Unet_v2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "exempt-chain",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-----------------------------------------\n",
      "keras-unet init: TF version is >= 2.0.0 - using `tf.keras` instead of `Keras`\n",
      "-----------------------------------------\n"
     ]
    }
   ],
   "source": [
    "from keras_unet.models import custom_unet\n",
    "from keras.optimizers import Adam, SGD\n",
    "from keras_unet.metrics import iou, iou_thresholded\n",
    "import tensorflow as tf \n",
    "from PIL import Image\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "gpus = tf.config.experimental.list_physical_devices('GPU') \n",
    "for gpu in gpus:\n",
    "    tf.config.experimental.set_memory_growth(gpu, True)\n",
    "\n",
    "IMG_LOC = '/home/clint/projects/DSP/data/p3/project3/data/'\n",
    "MASK_LOC = '/home/clint/projects/DSP/data/p3/project3/masks/'\n",
    "TRAIN_TXT = '/home/clint/projects/DSP/data/p3/project3/train.txt'\n",
    "TEST_TXT = '/home/clint/projects/DSP/data/p3/project3/test.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "virgin-columbia",
   "metadata": {},
   "outputs": [],
   "source": [
    "def img_preprocessing(img_loc, mask_loc):\n",
    "    pairs = []\n",
    "    with open(TRAIN_TXT) as train_file:\n",
    "        for line in train_file:\n",
    "            line = line.replace('\\n', '')\n",
    "            for im_file in os.listdir(f'{img_loc}/{line}/data/{line}'):\n",
    "                pairs.append([f'{img_loc}/{line}/data/{line}/{im_file}',f'{mask_loc}/{line}.png'])\n",
    "\n",
    "        return pairs\n",
    "pairs = img_preprocessing(IMG_LOC, MASK_LOC)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ordinary-executive",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pairs = pairs[:4000]\n",
    "val_pairs = pairs[:4000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "surrounded-fence",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.utils import to_categorical ,Sequence\n",
    "class DataGenerator(Sequence):\n",
    "    'Generates data for Keras'\n",
    "    \n",
    "    def __init__(self, pair, batch_size=16, dim=(224,224,3), shuffle=True):\n",
    "        'Initialization'\n",
    "        self.dim = dim\n",
    "        self.pair = pair\n",
    "#         self.class_map = class_map\n",
    "        self.batch_size = batch_size\n",
    "        self.shuffle = shuffle\n",
    "        self.on_epoch_end()\n",
    "\n",
    "    def __len__(self):\n",
    "        'Denotes the number of batches per epoch'\n",
    "        return int(np.floor(len(self.pair) / self.batch_size))\n",
    "\n",
    "    def __getitem__(self, index):\n",
    "        'Generate one batch of data'\n",
    "        # Generate indexes of the batch\n",
    "        indexes = self.indexes[index*self.batch_size:(index+1)*self.batch_size]\n",
    "\n",
    "        # Find list of IDs\n",
    "        list_IDs_temp = [k for k in indexes]\n",
    "\n",
    "        # Generate data\n",
    "        X, y = self.__data_generation(list_IDs_temp)\n",
    "\n",
    "        return X, y\n",
    "\n",
    "    def on_epoch_end(self):\n",
    "        'Updates indexes after each epoch'\n",
    "        self.indexes = np.arange(len(self.pair))\n",
    "        if self.shuffle == True:\n",
    "            np.random.shuffle(self.indexes)\n",
    "\n",
    "    def __data_generation(self, list_IDs_temp):\n",
    "        'Generates data containing batch_size samples' # X : (n_samples, *dim, n_channels)\n",
    "        # Initialization\n",
    "        batch_imgs = list()\n",
    "        batch_labels = list()\n",
    "\n",
    "        # Generate data\n",
    "        for i in list_IDs_temp:\n",
    "            # Store sample\n",
    "            img = Image.open(self.pair[i][0])\n",
    "            mask = Image.open(self.pair[i][1])\n",
    "            width, height = img.size   # Get dimensions\n",
    "            left = (width - 256)/2\n",
    "            top = (height - 256)/2\n",
    "            right = (width + 256)/2\n",
    "            bottom = (height + 256)/2\n",
    "            # Crop the center of the image\n",
    "            img = img.crop((left, top, right, bottom))\n",
    "            mask = mask.crop((left, top, right, bottom))\n",
    "            img = np.asarray(img, dtype=np.float32)/255\n",
    "            batch_imgs.append(img)\n",
    "            \n",
    "            mask = np.asarray(mask, dtype=np.float32)/2\n",
    "            batch_labels.append(mask)\n",
    "\n",
    "            \n",
    "        return np.array(batch_imgs) ,np.array(batch_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "unavailable-gender",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "500"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img_size = 256\n",
    "train_generator = DataGenerator(train_pairs,batch_size=8, dim=(img_size,img_size,1) ,shuffle=True)\n",
    "train_steps = train_generator.__len__()\n",
    "train_steps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "demanding-district",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8, 256, 256)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X,y = train_generator.__getitem__(1)\n",
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "compliant-facing",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "500"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_generator = DataGenerator(val_pairs,batch_size=8, dim=(img_size,img_size,1) ,shuffle=True)\n",
    "val_steps = val_generator.__len__()\n",
    "val_steps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "czech-bundle",
   "metadata": {},
   "outputs": [],
   "source": [
    "def conv_block(tensor, nfilters, size=3, padding='same', initializer=\"he_normal\"):\n",
    "    x = Conv2D(filters=nfilters, kernel_size=(size, size), padding=padding, kernel_initializer=initializer)(tensor)\n",
    "    x = BatchNormalization()(x)\n",
    "    x = Activation(\"relu\")(x)\n",
    "    x = Conv2D(filters=nfilters, kernel_size=(size, size), padding=padding, kernel_initializer=initializer)(x)\n",
    "    x = BatchNormalization()(x)\n",
    "    x = Activation(\"relu\")(x)\n",
    "    return x\n",
    "\n",
    "\n",
    "def deconv_block(tensor, residual, nfilters, size=3, padding='same', strides=(2, 2)):\n",
    "    y = Conv2DTranspose(nfilters, kernel_size=(size, size), strides=strides, padding=padding)(tensor)\n",
    "    y = concatenate([y, residual], axis=3)\n",
    "    y = conv_block(y, nfilters)\n",
    "    return y\n",
    "\n",
    "\n",
    "def Unet(h, w, filters):\n",
    "    input_layer = Input(shape=(h, w, 1), name='image_input')\n",
    "    conv1 = conv_block(input_layer, nfilters=filters)\n",
    "    conv1_out = MaxPooling2D(pool_size=(2, 2))(conv1)\n",
    "    conv2 = conv_block(conv1_out, nfilters=filters*2)\n",
    "    conv2_out = MaxPooling2D(pool_size=(2, 2))(conv2)\n",
    "    conv3 = conv_block(conv2_out, nfilters=filters*4)\n",
    "    conv3_out = MaxPooling2D(pool_size=(2, 2))(conv3)\n",
    "    conv4 = conv_block(conv3_out, nfilters=filters*8)\n",
    "    conv4_out = MaxPooling2D(pool_size=(2, 2))(conv4)\n",
    "    conv4_out = Dropout(0.5)(conv4_out)\n",
    "    conv5 = conv_block(conv4_out, nfilters=filters*16)\n",
    "    conv5 = Dropout(0.5)(conv5)\n",
    "\n",
    "    deconv6 = deconv_block(conv5, residual=conv4, nfilters=filters*8)\n",
    "    deconv6 = Dropout(0.5)(deconv6)\n",
    "    deconv7 = deconv_block(deconv6, residual=conv3, nfilters=filters*4)\n",
    "    deconv7 = Dropout(0.5)(deconv7) \n",
    "    deconv8 = deconv_block(deconv7, residual=conv2, nfilters=filters*2)\n",
    "    deconv9 = deconv_block(deconv8, residual=conv1, nfilters=filters)\n",
    "    output_layer = Conv2D(filters=1, kernel_size=(1, 1), activation='softmax')(deconv9)\n",
    "\n",
    "    model = Model(inputs=input_layer, outputs=output_layer, name='Unet')\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "immediate-california",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
    "from matplotlib import pyplot as plt\n",
    "import cv2\n",
    "from tensorflow.keras.preprocessing.image import load_img, img_to_array\n",
    "from tensorflow.keras.utils import to_categorical ,Sequence\n",
    "from tensorflow.keras import backend as K\n",
    "from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, concatenate, Conv2DTranspose, BatchNormalization, Activation, Dropout\n",
    "from tensorflow.keras.optimizers import Adadelta, Nadam ,Adam\n",
    "from tensorflow.keras.models import Model, load_model\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau, CSVLogger, TensorBoard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "binary-orientation",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"Unet\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "image_input (InputLayer)        [(None, 256, 256, 1) 0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_57 (Conv2D)              (None, 256, 256, 1)  10          image_input[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_54 (BatchNo (None, 256, 256, 1)  4           conv2d_57[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_54 (Activation)      (None, 256, 256, 1)  0           batch_normalization_54[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_58 (Conv2D)              (None, 256, 256, 1)  10          activation_54[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_55 (BatchNo (None, 256, 256, 1)  4           conv2d_58[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_55 (Activation)      (None, 256, 256, 1)  0           batch_normalization_55[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_12 (MaxPooling2D) (None, 128, 128, 1)  0           activation_55[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_59 (Conv2D)              (None, 128, 128, 2)  20          max_pooling2d_12[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_56 (BatchNo (None, 128, 128, 2)  8           conv2d_59[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_56 (Activation)      (None, 128, 128, 2)  0           batch_normalization_56[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_60 (Conv2D)              (None, 128, 128, 2)  38          activation_56[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_57 (BatchNo (None, 128, 128, 2)  8           conv2d_60[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_57 (Activation)      (None, 128, 128, 2)  0           batch_normalization_57[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_13 (MaxPooling2D) (None, 64, 64, 2)    0           activation_57[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_61 (Conv2D)              (None, 64, 64, 4)    76          max_pooling2d_13[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_58 (BatchNo (None, 64, 64, 4)    16          conv2d_61[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_58 (Activation)      (None, 64, 64, 4)    0           batch_normalization_58[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_62 (Conv2D)              (None, 64, 64, 4)    148         activation_58[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_59 (BatchNo (None, 64, 64, 4)    16          conv2d_62[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_59 (Activation)      (None, 64, 64, 4)    0           batch_normalization_59[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_14 (MaxPooling2D) (None, 32, 32, 4)    0           activation_59[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_63 (Conv2D)              (None, 32, 32, 8)    296         max_pooling2d_14[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_60 (BatchNo (None, 32, 32, 8)    32          conv2d_63[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_60 (Activation)      (None, 32, 32, 8)    0           batch_normalization_60[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_64 (Conv2D)              (None, 32, 32, 8)    584         activation_60[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_61 (BatchNo (None, 32, 32, 8)    32          conv2d_64[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_61 (Activation)      (None, 32, 32, 8)    0           batch_normalization_61[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_15 (MaxPooling2D) (None, 16, 16, 8)    0           activation_61[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "dropout_12 (Dropout)            (None, 16, 16, 8)    0           max_pooling2d_15[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_65 (Conv2D)              (None, 16, 16, 16)   1168        dropout_12[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_62 (BatchNo (None, 16, 16, 16)   64          conv2d_65[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_62 (Activation)      (None, 16, 16, 16)   0           batch_normalization_62[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_66 (Conv2D)              (None, 16, 16, 16)   2320        activation_62[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_63 (BatchNo (None, 16, 16, 16)   64          conv2d_66[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_63 (Activation)      (None, 16, 16, 16)   0           batch_normalization_63[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "dropout_13 (Dropout)            (None, 16, 16, 16)   0           activation_63[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_12 (Conv2DTran (None, 32, 32, 8)    1160        dropout_13[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_12 (Concatenate)    (None, 32, 32, 16)   0           conv2d_transpose_12[0][0]        \n",
      "                                                                 activation_61[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_67 (Conv2D)              (None, 32, 32, 8)    1160        concatenate_12[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_64 (BatchNo (None, 32, 32, 8)    32          conv2d_67[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_64 (Activation)      (None, 32, 32, 8)    0           batch_normalization_64[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_68 (Conv2D)              (None, 32, 32, 8)    584         activation_64[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_65 (BatchNo (None, 32, 32, 8)    32          conv2d_68[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_65 (Activation)      (None, 32, 32, 8)    0           batch_normalization_65[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "dropout_14 (Dropout)            (None, 32, 32, 8)    0           activation_65[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_13 (Conv2DTran (None, 64, 64, 4)    292         dropout_14[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_13 (Concatenate)    (None, 64, 64, 8)    0           conv2d_transpose_13[0][0]        \n",
      "                                                                 activation_59[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_69 (Conv2D)              (None, 64, 64, 4)    292         concatenate_13[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_66 (BatchNo (None, 64, 64, 4)    16          conv2d_69[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_66 (Activation)      (None, 64, 64, 4)    0           batch_normalization_66[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_70 (Conv2D)              (None, 64, 64, 4)    148         activation_66[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_67 (BatchNo (None, 64, 64, 4)    16          conv2d_70[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_67 (Activation)      (None, 64, 64, 4)    0           batch_normalization_67[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "dropout_15 (Dropout)            (None, 64, 64, 4)    0           activation_67[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_14 (Conv2DTran (None, 128, 128, 2)  74          dropout_15[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_14 (Concatenate)    (None, 128, 128, 4)  0           conv2d_transpose_14[0][0]        \n",
      "                                                                 activation_57[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_71 (Conv2D)              (None, 128, 128, 2)  74          concatenate_14[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_68 (BatchNo (None, 128, 128, 2)  8           conv2d_71[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_68 (Activation)      (None, 128, 128, 2)  0           batch_normalization_68[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_72 (Conv2D)              (None, 128, 128, 2)  38          activation_68[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_69 (BatchNo (None, 128, 128, 2)  8           conv2d_72[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_69 (Activation)      (None, 128, 128, 2)  0           batch_normalization_69[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_15 (Conv2DTran (None, 256, 256, 1)  19          activation_69[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_15 (Concatenate)    (None, 256, 256, 2)  0           conv2d_transpose_15[0][0]        \n",
      "                                                                 activation_55[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_73 (Conv2D)              (None, 256, 256, 1)  19          concatenate_15[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_70 (BatchNo (None, 256, 256, 1)  4           conv2d_73[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_70 (Activation)      (None, 256, 256, 1)  0           batch_normalization_70[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_74 (Conv2D)              (None, 256, 256, 1)  10          activation_70[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_71 (BatchNo (None, 256, 256, 1)  4           conv2d_74[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_71 (Activation)      (None, 256, 256, 1)  0           batch_normalization_71[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_75 (Conv2D)              (None, 256, 256, 1)  2           activation_71[0][0]              \n",
      "==================================================================================================\n",
      "Total params: 8,910\n",
      "Trainable params: 8,726\n",
      "Non-trainable params: 184\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Unet(img_size , img_size , 1)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "monetary-recruitment",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "neither-study",
   "metadata": {},
   "outputs": [],
   "source": [
    "m = ModelCheckpoint(mode='max', filepath='top-weights.h5', monitor='val_acc',save_best_only='True', save_weights_only='True', verbose=1)\n",
    "e = EarlyStopping(mode='max', monitor='val_accuracy', patience=10, verbose=0)\n",
    "t = TensorBoard(log_dir=\"logs/\", histogram_freq=0, write_graph=True, write_images=False)\n",
    "r = ReduceLROnPlateau(monitor='val_accuracy',factor=0.1,patience=5,verbose=1,mode=\"max\",min_lr=0.001)\n",
    "c = CSVLogger(\"logs/log.csv\" , append=True , separator=',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "single-portland",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "332/500 [==================>...........] - ETA: 52s - loss: 494.5722 - accuracy: 0.1050"
     ]
    }
   ],
   "source": [
    "results = model.fit_generator(train_generator, \n",
    "                              epochs=20,\n",
    "                              steps_per_epoch=train_steps,\n",
    "                              validation_steps=val_steps,\n",
    "                              validation_data=val_generator,\n",
    "                              callbacks=[m,e,t,r,c])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "alien-cedar",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "broken-conservation",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "classified-mirror",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "failing-balloon",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "third-layer",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "english-auction",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fossil-binary",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "billion-steel",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "recognized-reynolds",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "atlantic-vietnam",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "flush-bankruptcy",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "alone-explorer",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "apparent-geometry",
   "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": "3.6.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}