train.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Dependency libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from model.danet import danet_resnet101\n",
    "from utils.loss import *\n",
    "from utils.metrics import *\n",
    "from dataloaders.generater import *\n",
    "import os\n",
    "from keras.callbacks import ModelCheckpoint, EarlyStopping\n",
    "from keras.optimizers import SGD\n",
    "from keras.callbacks import ReduceLROnPlateau\n",
    "from keras.utils.training_utils import multi_gpu_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# network params\n",
    "BatchSize = 5\n",
    "NumChannels = 3\n",
    "ImgHeight = 512\n",
    "ImgWidth = 512\n",
    "NumClass = 1\n",
    "\n",
    "# training params\n",
    "GPUs = '0, 1, 2, 3'\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = GPUs\n",
    "Optimizer = 'Adam'  # SGD(lr=0.01, momentum=0.9, nesterov=True)\n",
    "Loss = ce_jaccard_loss\n",
    "Metrics = ['accuracy', iou]\n",
    "NumEpochs = 100\n",
    "Patience = 10\n",
    "\n",
    "# data params\n",
    "TrainImageDir = '/data/dh_zhu/cong/AerialImageDataset/train/images/'\n",
    "ValImageDir = '/data/dh_zhu/cong/AerialImageDataset/val/images/'\n",
    "\n",
    "# visualization params\n",
    "metric_list = ['acc', 'iou']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 512, 512, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)               (None, 256, 256, 64) 9408        input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_1 (BatchNor (None, 256, 256, 64) 256         conv2d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_1 (Activation)       (None, 256, 256, 64) 0           batch_normalization_1[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)  (None, 128, 128, 64) 0           activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)               (None, 128, 128, 64) 4096        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_2 (BatchNor (None, 128, 128, 64) 256         conv2d_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_2 (Activation)       (None, 128, 128, 64) 0           batch_normalization_2[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)               (None, 128, 128, 64) 36864       activation_2[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_3 (BatchNor (None, 128, 128, 64) 256         conv2d_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_3 (Activation)       (None, 128, 128, 64) 0           batch_normalization_3[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)               (None, 128, 128, 256 16384       activation_3[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_5 (Conv2D)               (None, 128, 128, 256 16384       max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_4 (BatchNor (None, 128, 128, 256 1024        conv2d_4[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_5 (BatchNor (None, 128, 128, 256 1024        conv2d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "add_1 (Add)                     (None, 128, 128, 256 0           batch_normalization_4[0][0]      \n",
      "                                                                 batch_normalization_5[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_4 (Activation)       (None, 128, 128, 256 0           add_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_6 (Conv2D)               (None, 128, 128, 64) 16384       activation_4[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_6 (BatchNor (None, 128, 128, 64) 256         conv2d_6[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_5 (Activation)       (None, 128, 128, 64) 0           batch_normalization_6[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_7 (Conv2D)               (None, 128, 128, 64) 36864       activation_5[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_7 (BatchNor (None, 128, 128, 64) 256         conv2d_7[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_6 (Activation)       (None, 128, 128, 64) 0           batch_normalization_7[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_8 (Conv2D)               (None, 128, 128, 256 16384       activation_6[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_8 (BatchNor (None, 128, 128, 256 1024        conv2d_8[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "add_2 (Add)                     (None, 128, 128, 256 0           batch_normalization_8[0][0]      \n",
      "                                                                 activation_4[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_7 (Activation)       (None, 128, 128, 256 0           add_2[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_9 (Conv2D)               (None, 128, 128, 64) 16384       activation_7[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_9 (BatchNor (None, 128, 128, 64) 256         conv2d_9[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_8 (Activation)       (None, 128, 128, 64) 0           batch_normalization_9[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_10 (Conv2D)              (None, 128, 128, 64) 36864       activation_8[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_10 (BatchNo (None, 128, 128, 64) 256         conv2d_10[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_9 (Activation)       (None, 128, 128, 64) 0           batch_normalization_10[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_11 (Conv2D)              (None, 128, 128, 256 16384       activation_9[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_11 (BatchNo (None, 128, 128, 256 1024        conv2d_11[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_3 (Add)                     (None, 128, 128, 256 0           batch_normalization_11[0][0]     \n",
      "                                                                 activation_7[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_10 (Activation)      (None, 128, 128, 256 0           add_3[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_12 (Conv2D)              (None, 128, 128, 128 32768       activation_10[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_12 (BatchNo (None, 128, 128, 128 512         conv2d_12[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_11 (Activation)      (None, 128, 128, 128 0           batch_normalization_12[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_13 (Conv2D)              (None, 64, 64, 128)  147456      activation_11[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_13 (BatchNo (None, 64, 64, 128)  512         conv2d_13[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_12 (Activation)      (None, 64, 64, 128)  0           batch_normalization_13[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_14 (Conv2D)              (None, 64, 64, 512)  65536       activation_12[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_15 (Conv2D)              (None, 64, 64, 512)  131072      activation_10[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_14 (BatchNo (None, 64, 64, 512)  2048        conv2d_14[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_15 (BatchNo (None, 64, 64, 512)  2048        conv2d_15[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_4 (Add)                     (None, 64, 64, 512)  0           batch_normalization_14[0][0]     \n",
      "                                                                 batch_normalization_15[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_13 (Activation)      (None, 64, 64, 512)  0           add_4[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_16 (Conv2D)              (None, 64, 64, 128)  65536       activation_13[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_16 (BatchNo (None, 64, 64, 128)  512         conv2d_16[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_14 (Activation)      (None, 64, 64, 128)  0           batch_normalization_16[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_17 (Conv2D)              (None, 64, 64, 128)  147456      activation_14[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_17 (BatchNo (None, 64, 64, 128)  512         conv2d_17[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_15 (Activation)      (None, 64, 64, 128)  0           batch_normalization_17[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_18 (Conv2D)              (None, 64, 64, 512)  65536       activation_15[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_18 (BatchNo (None, 64, 64, 512)  2048        conv2d_18[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_5 (Add)                     (None, 64, 64, 512)  0           batch_normalization_18[0][0]     \n",
      "                                                                 activation_13[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_16 (Activation)      (None, 64, 64, 512)  0           add_5[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_19 (Conv2D)              (None, 64, 64, 128)  65536       activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_19 (BatchNo (None, 64, 64, 128)  512         conv2d_19[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_17 (Activation)      (None, 64, 64, 128)  0           batch_normalization_19[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_20 (Conv2D)              (None, 64, 64, 128)  147456      activation_17[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_20 (BatchNo (None, 64, 64, 128)  512         conv2d_20[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_18 (Activation)      (None, 64, 64, 128)  0           batch_normalization_20[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_21 (Conv2D)              (None, 64, 64, 512)  65536       activation_18[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_21 (BatchNo (None, 64, 64, 512)  2048        conv2d_21[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_6 (Add)                     (None, 64, 64, 512)  0           batch_normalization_21[0][0]     \n",
      "                                                                 activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_19 (Activation)      (None, 64, 64, 512)  0           add_6[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_22 (Conv2D)              (None, 64, 64, 128)  65536       activation_19[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_22 (BatchNo (None, 64, 64, 128)  512         conv2d_22[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_20 (Activation)      (None, 64, 64, 128)  0           batch_normalization_22[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_23 (Conv2D)              (None, 64, 64, 128)  147456      activation_20[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_23 (BatchNo (None, 64, 64, 128)  512         conv2d_23[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_21 (Activation)      (None, 64, 64, 128)  0           batch_normalization_23[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_24 (Conv2D)              (None, 64, 64, 512)  65536       activation_21[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_24 (BatchNo (None, 64, 64, 512)  2048        conv2d_24[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_7 (Add)                     (None, 64, 64, 512)  0           batch_normalization_24[0][0]     \n",
      "                                                                 activation_19[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_22 (Activation)      (None, 64, 64, 512)  0           add_7[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_25 (Conv2D)              (None, 64, 64, 256)  131072      activation_22[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_25 (BatchNo (None, 64, 64, 256)  1024        conv2d_25[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_23 (Activation)      (None, 64, 64, 256)  0           batch_normalization_25[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_26 (Conv2D)              (None, 64, 64, 256)  589824      activation_23[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_26 (BatchNo (None, 64, 64, 256)  1024        conv2d_26[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_24 (Activation)      (None, 64, 64, 256)  0           batch_normalization_26[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_27 (Conv2D)              (None, 64, 64, 1024) 262144      activation_24[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_28 (Conv2D)              (None, 64, 64, 1024) 524288      activation_22[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_27 (BatchNo (None, 64, 64, 1024) 4096        conv2d_27[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_28 (BatchNo (None, 64, 64, 1024) 4096        conv2d_28[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_8 (Add)                     (None, 64, 64, 1024) 0           batch_normalization_27[0][0]     \n",
      "                                                                 batch_normalization_28[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_25 (Activation)      (None, 64, 64, 1024) 0           add_8[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_29 (Conv2D)              (None, 64, 64, 256)  262144      activation_25[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_29 (BatchNo (None, 64, 64, 256)  1024        conv2d_29[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_26 (Activation)      (None, 64, 64, 256)  0           batch_normalization_29[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_30 (Conv2D)              (None, 64, 64, 256)  589824      activation_26[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_30 (BatchNo (None, 64, 64, 256)  1024        conv2d_30[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_27 (Activation)      (None, 64, 64, 256)  0           batch_normalization_30[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_31 (Conv2D)              (None, 64, 64, 1024) 262144      activation_27[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_31 (BatchNo (None, 64, 64, 1024) 4096        conv2d_31[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_9 (Add)                     (None, 64, 64, 1024) 0           batch_normalization_31[0][0]     \n",
      "                                                                 activation_25[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_28 (Activation)      (None, 64, 64, 1024) 0           add_9[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_32 (Conv2D)              (None, 64, 64, 256)  262144      activation_28[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_32 (BatchNo (None, 64, 64, 256)  1024        conv2d_32[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_29 (Activation)      (None, 64, 64, 256)  0           batch_normalization_32[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_33 (Conv2D)              (None, 64, 64, 256)  589824      activation_29[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_33 (BatchNo (None, 64, 64, 256)  1024        conv2d_33[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_30 (Activation)      (None, 64, 64, 256)  0           batch_normalization_33[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_34 (Conv2D)              (None, 64, 64, 1024) 262144      activation_30[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_34 (BatchNo (None, 64, 64, 1024) 4096        conv2d_34[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_10 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_34[0][0]     \n",
      "                                                                 activation_28[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_31 (Activation)      (None, 64, 64, 1024) 0           add_10[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_35 (Conv2D)              (None, 64, 64, 256)  262144      activation_31[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_35 (BatchNo (None, 64, 64, 256)  1024        conv2d_35[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_32 (Activation)      (None, 64, 64, 256)  0           batch_normalization_35[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_36 (Conv2D)              (None, 64, 64, 256)  589824      activation_32[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_36 (BatchNo (None, 64, 64, 256)  1024        conv2d_36[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_33 (Activation)      (None, 64, 64, 256)  0           batch_normalization_36[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_37 (Conv2D)              (None, 64, 64, 1024) 262144      activation_33[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_37 (BatchNo (None, 64, 64, 1024) 4096        conv2d_37[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_11 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_37[0][0]     \n",
      "                                                                 activation_31[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_34 (Activation)      (None, 64, 64, 1024) 0           add_11[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_38 (Conv2D)              (None, 64, 64, 256)  262144      activation_34[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_38 (BatchNo (None, 64, 64, 256)  1024        conv2d_38[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_35 (Activation)      (None, 64, 64, 256)  0           batch_normalization_38[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_39 (Conv2D)              (None, 64, 64, 256)  589824      activation_35[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_39 (BatchNo (None, 64, 64, 256)  1024        conv2d_39[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_36 (Activation)      (None, 64, 64, 256)  0           batch_normalization_39[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_40 (Conv2D)              (None, 64, 64, 1024) 262144      activation_36[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_40 (BatchNo (None, 64, 64, 1024) 4096        conv2d_40[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_12 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_40[0][0]     \n",
      "                                                                 activation_34[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_37 (Activation)      (None, 64, 64, 1024) 0           add_12[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_41 (Conv2D)              (None, 64, 64, 256)  262144      activation_37[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_41 (BatchNo (None, 64, 64, 256)  1024        conv2d_41[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_38 (Activation)      (None, 64, 64, 256)  0           batch_normalization_41[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_42 (Conv2D)              (None, 64, 64, 256)  589824      activation_38[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_42 (BatchNo (None, 64, 64, 256)  1024        conv2d_42[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_39 (Activation)      (None, 64, 64, 256)  0           batch_normalization_42[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_43 (Conv2D)              (None, 64, 64, 1024) 262144      activation_39[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_43 (BatchNo (None, 64, 64, 1024) 4096        conv2d_43[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_13 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_43[0][0]     \n",
      "                                                                 activation_37[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_40 (Activation)      (None, 64, 64, 1024) 0           add_13[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_44 (Conv2D)              (None, 64, 64, 256)  262144      activation_40[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_44 (BatchNo (None, 64, 64, 256)  1024        conv2d_44[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_41 (Activation)      (None, 64, 64, 256)  0           batch_normalization_44[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_45 (Conv2D)              (None, 64, 64, 256)  589824      activation_41[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_45 (BatchNo (None, 64, 64, 256)  1024        conv2d_45[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_42 (Activation)      (None, 64, 64, 256)  0           batch_normalization_45[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_46 (Conv2D)              (None, 64, 64, 1024) 262144      activation_42[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_46 (BatchNo (None, 64, 64, 1024) 4096        conv2d_46[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_14 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_46[0][0]     \n",
      "                                                                 activation_40[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_43 (Activation)      (None, 64, 64, 1024) 0           add_14[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_47 (Conv2D)              (None, 64, 64, 256)  262144      activation_43[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_47 (BatchNo (None, 64, 64, 256)  1024        conv2d_47[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_44 (Activation)      (None, 64, 64, 256)  0           batch_normalization_47[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_48 (Conv2D)              (None, 64, 64, 256)  589824      activation_44[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_48 (BatchNo (None, 64, 64, 256)  1024        conv2d_48[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_45 (Activation)      (None, 64, 64, 256)  0           batch_normalization_48[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_49 (Conv2D)              (None, 64, 64, 1024) 262144      activation_45[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_49 (BatchNo (None, 64, 64, 1024) 4096        conv2d_49[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_15 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_49[0][0]     \n",
      "                                                                 activation_43[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_46 (Activation)      (None, 64, 64, 1024) 0           add_15[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_50 (Conv2D)              (None, 64, 64, 256)  262144      activation_46[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_50 (BatchNo (None, 64, 64, 256)  1024        conv2d_50[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_47 (Activation)      (None, 64, 64, 256)  0           batch_normalization_50[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_51 (Conv2D)              (None, 64, 64, 256)  589824      activation_47[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_51 (BatchNo (None, 64, 64, 256)  1024        conv2d_51[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_48 (Activation)      (None, 64, 64, 256)  0           batch_normalization_51[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_52 (Conv2D)              (None, 64, 64, 1024) 262144      activation_48[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_52 (BatchNo (None, 64, 64, 1024) 4096        conv2d_52[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_16 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_52[0][0]     \n",
      "                                                                 activation_46[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_49 (Activation)      (None, 64, 64, 1024) 0           add_16[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_53 (Conv2D)              (None, 64, 64, 256)  262144      activation_49[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_53 (BatchNo (None, 64, 64, 256)  1024        conv2d_53[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_50 (Activation)      (None, 64, 64, 256)  0           batch_normalization_53[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_54 (Conv2D)              (None, 64, 64, 256)  589824      activation_50[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_54 (BatchNo (None, 64, 64, 256)  1024        conv2d_54[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_51 (Activation)      (None, 64, 64, 256)  0           batch_normalization_54[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_55 (Conv2D)              (None, 64, 64, 1024) 262144      activation_51[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_55 (BatchNo (None, 64, 64, 1024) 4096        conv2d_55[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_17 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_55[0][0]     \n",
      "                                                                 activation_49[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_52 (Activation)      (None, 64, 64, 1024) 0           add_17[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_56 (Conv2D)              (None, 64, 64, 256)  262144      activation_52[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_56 (BatchNo (None, 64, 64, 256)  1024        conv2d_56[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_53 (Activation)      (None, 64, 64, 256)  0           batch_normalization_56[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_57 (Conv2D)              (None, 64, 64, 256)  589824      activation_53[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_57 (BatchNo (None, 64, 64, 256)  1024        conv2d_57[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_54 (Activation)      (None, 64, 64, 256)  0           batch_normalization_57[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_58 (Conv2D)              (None, 64, 64, 1024) 262144      activation_54[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_58 (BatchNo (None, 64, 64, 1024) 4096        conv2d_58[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_18 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_58[0][0]     \n",
      "                                                                 activation_52[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_55 (Activation)      (None, 64, 64, 1024) 0           add_18[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_59 (Conv2D)              (None, 64, 64, 256)  262144      activation_55[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_59 (BatchNo (None, 64, 64, 256)  1024        conv2d_59[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_56 (Activation)      (None, 64, 64, 256)  0           batch_normalization_59[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_60 (Conv2D)              (None, 64, 64, 256)  589824      activation_56[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_60 (BatchNo (None, 64, 64, 256)  1024        conv2d_60[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_57 (Activation)      (None, 64, 64, 256)  0           batch_normalization_60[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_61 (Conv2D)              (None, 64, 64, 1024) 262144      activation_57[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_61 (BatchNo (None, 64, 64, 1024) 4096        conv2d_61[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_19 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_61[0][0]     \n",
      "                                                                 activation_55[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_58 (Activation)      (None, 64, 64, 1024) 0           add_19[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_62 (Conv2D)              (None, 64, 64, 256)  262144      activation_58[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_62 (BatchNo (None, 64, 64, 256)  1024        conv2d_62[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_59 (Activation)      (None, 64, 64, 256)  0           batch_normalization_62[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_63 (Conv2D)              (None, 64, 64, 256)  589824      activation_59[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_63 (BatchNo (None, 64, 64, 256)  1024        conv2d_63[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_60 (Activation)      (None, 64, 64, 256)  0           batch_normalization_63[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_64 (Conv2D)              (None, 64, 64, 1024) 262144      activation_60[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_64 (BatchNo (None, 64, 64, 1024) 4096        conv2d_64[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_20 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_64[0][0]     \n",
      "                                                                 activation_58[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_61 (Activation)      (None, 64, 64, 1024) 0           add_20[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_65 (Conv2D)              (None, 64, 64, 256)  262144      activation_61[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_65 (BatchNo (None, 64, 64, 256)  1024        conv2d_65[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_62 (Activation)      (None, 64, 64, 256)  0           batch_normalization_65[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_66 (Conv2D)              (None, 64, 64, 256)  589824      activation_62[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_66 (BatchNo (None, 64, 64, 256)  1024        conv2d_66[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_63 (Activation)      (None, 64, 64, 256)  0           batch_normalization_66[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_67 (Conv2D)              (None, 64, 64, 1024) 262144      activation_63[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_67 (BatchNo (None, 64, 64, 1024) 4096        conv2d_67[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_21 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_67[0][0]     \n",
      "                                                                 activation_61[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_64 (Activation)      (None, 64, 64, 1024) 0           add_21[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_68 (Conv2D)              (None, 64, 64, 256)  262144      activation_64[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_68 (BatchNo (None, 64, 64, 256)  1024        conv2d_68[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_65 (Activation)      (None, 64, 64, 256)  0           batch_normalization_68[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_69 (Conv2D)              (None, 64, 64, 256)  589824      activation_65[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_69 (BatchNo (None, 64, 64, 256)  1024        conv2d_69[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_66 (Activation)      (None, 64, 64, 256)  0           batch_normalization_69[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_70 (Conv2D)              (None, 64, 64, 1024) 262144      activation_66[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_70 (BatchNo (None, 64, 64, 1024) 4096        conv2d_70[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_22 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_70[0][0]     \n",
      "                                                                 activation_64[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_67 (Activation)      (None, 64, 64, 1024) 0           add_22[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_71 (Conv2D)              (None, 64, 64, 256)  262144      activation_67[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_71 (BatchNo (None, 64, 64, 256)  1024        conv2d_71[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_68 (Activation)      (None, 64, 64, 256)  0           batch_normalization_71[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_72 (Conv2D)              (None, 64, 64, 256)  589824      activation_68[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_72 (BatchNo (None, 64, 64, 256)  1024        conv2d_72[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_69 (Activation)      (None, 64, 64, 256)  0           batch_normalization_72[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_73 (Conv2D)              (None, 64, 64, 1024) 262144      activation_69[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_73 (BatchNo (None, 64, 64, 1024) 4096        conv2d_73[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_23 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_73[0][0]     \n",
      "                                                                 activation_67[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_70 (Activation)      (None, 64, 64, 1024) 0           add_23[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_74 (Conv2D)              (None, 64, 64, 256)  262144      activation_70[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_74 (BatchNo (None, 64, 64, 256)  1024        conv2d_74[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_71 (Activation)      (None, 64, 64, 256)  0           batch_normalization_74[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_75 (Conv2D)              (None, 64, 64, 256)  589824      activation_71[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_75 (BatchNo (None, 64, 64, 256)  1024        conv2d_75[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_72 (Activation)      (None, 64, 64, 256)  0           batch_normalization_75[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_76 (Conv2D)              (None, 64, 64, 1024) 262144      activation_72[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_76 (BatchNo (None, 64, 64, 1024) 4096        conv2d_76[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_24 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_76[0][0]     \n",
      "                                                                 activation_70[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_73 (Activation)      (None, 64, 64, 1024) 0           add_24[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_77 (Conv2D)              (None, 64, 64, 256)  262144      activation_73[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_77 (BatchNo (None, 64, 64, 256)  1024        conv2d_77[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_74 (Activation)      (None, 64, 64, 256)  0           batch_normalization_77[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_78 (Conv2D)              (None, 64, 64, 256)  589824      activation_74[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_78 (BatchNo (None, 64, 64, 256)  1024        conv2d_78[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_75 (Activation)      (None, 64, 64, 256)  0           batch_normalization_78[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_79 (Conv2D)              (None, 64, 64, 1024) 262144      activation_75[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_79 (BatchNo (None, 64, 64, 1024) 4096        conv2d_79[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_25 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_79[0][0]     \n",
      "                                                                 activation_73[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_76 (Activation)      (None, 64, 64, 1024) 0           add_25[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_80 (Conv2D)              (None, 64, 64, 256)  262144      activation_76[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_80 (BatchNo (None, 64, 64, 256)  1024        conv2d_80[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_77 (Activation)      (None, 64, 64, 256)  0           batch_normalization_80[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_81 (Conv2D)              (None, 64, 64, 256)  589824      activation_77[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_81 (BatchNo (None, 64, 64, 256)  1024        conv2d_81[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_78 (Activation)      (None, 64, 64, 256)  0           batch_normalization_81[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_82 (Conv2D)              (None, 64, 64, 1024) 262144      activation_78[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_82 (BatchNo (None, 64, 64, 1024) 4096        conv2d_82[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_26 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_82[0][0]     \n",
      "                                                                 activation_76[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_79 (Activation)      (None, 64, 64, 1024) 0           add_26[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_83 (Conv2D)              (None, 64, 64, 256)  262144      activation_79[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_83 (BatchNo (None, 64, 64, 256)  1024        conv2d_83[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_80 (Activation)      (None, 64, 64, 256)  0           batch_normalization_83[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_84 (Conv2D)              (None, 64, 64, 256)  589824      activation_80[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_84 (BatchNo (None, 64, 64, 256)  1024        conv2d_84[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_81 (Activation)      (None, 64, 64, 256)  0           batch_normalization_84[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_85 (Conv2D)              (None, 64, 64, 1024) 262144      activation_81[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_85 (BatchNo (None, 64, 64, 1024) 4096        conv2d_85[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_27 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_85[0][0]     \n",
      "                                                                 activation_79[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_82 (Activation)      (None, 64, 64, 1024) 0           add_27[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_86 (Conv2D)              (None, 64, 64, 256)  262144      activation_82[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_86 (BatchNo (None, 64, 64, 256)  1024        conv2d_86[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_83 (Activation)      (None, 64, 64, 256)  0           batch_normalization_86[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_87 (Conv2D)              (None, 64, 64, 256)  589824      activation_83[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_87 (BatchNo (None, 64, 64, 256)  1024        conv2d_87[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_84 (Activation)      (None, 64, 64, 256)  0           batch_normalization_87[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_88 (Conv2D)              (None, 64, 64, 1024) 262144      activation_84[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_88 (BatchNo (None, 64, 64, 1024) 4096        conv2d_88[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_28 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_88[0][0]     \n",
      "                                                                 activation_82[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_85 (Activation)      (None, 64, 64, 1024) 0           add_28[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_89 (Conv2D)              (None, 64, 64, 256)  262144      activation_85[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_89 (BatchNo (None, 64, 64, 256)  1024        conv2d_89[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_86 (Activation)      (None, 64, 64, 256)  0           batch_normalization_89[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_90 (Conv2D)              (None, 64, 64, 256)  589824      activation_86[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_90 (BatchNo (None, 64, 64, 256)  1024        conv2d_90[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_87 (Activation)      (None, 64, 64, 256)  0           batch_normalization_90[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_91 (Conv2D)              (None, 64, 64, 1024) 262144      activation_87[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_91 (BatchNo (None, 64, 64, 1024) 4096        conv2d_91[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_29 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_91[0][0]     \n",
      "                                                                 activation_85[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_88 (Activation)      (None, 64, 64, 1024) 0           add_29[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_92 (Conv2D)              (None, 64, 64, 256)  262144      activation_88[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_92 (BatchNo (None, 64, 64, 256)  1024        conv2d_92[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_89 (Activation)      (None, 64, 64, 256)  0           batch_normalization_92[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_93 (Conv2D)              (None, 64, 64, 256)  589824      activation_89[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_93 (BatchNo (None, 64, 64, 256)  1024        conv2d_93[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_90 (Activation)      (None, 64, 64, 256)  0           batch_normalization_93[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_94 (Conv2D)              (None, 64, 64, 1024) 262144      activation_90[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_94 (BatchNo (None, 64, 64, 1024) 4096        conv2d_94[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_30 (Add)                    (None, 64, 64, 1024) 0           batch_normalization_94[0][0]     \n",
      "                                                                 activation_88[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_91 (Activation)      (None, 64, 64, 1024) 0           add_30[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_95 (Conv2D)              (None, 64, 64, 512)  524288      activation_91[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_95 (BatchNo (None, 64, 64, 512)  2048        conv2d_95[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_92 (Activation)      (None, 64, 64, 512)  0           batch_normalization_95[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_96 (Conv2D)              (None, 64, 64, 512)  2359296     activation_92[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_96 (BatchNo (None, 64, 64, 512)  2048        conv2d_96[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_93 (Activation)      (None, 64, 64, 512)  0           batch_normalization_96[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_97 (Conv2D)              (None, 64, 64, 2048) 1048576     activation_93[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_98 (Conv2D)              (None, 64, 64, 2048) 2097152     activation_91[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_97 (BatchNo (None, 64, 64, 2048) 8192        conv2d_97[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_98 (BatchNo (None, 64, 64, 2048) 8192        conv2d_98[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_31 (Add)                    (None, 64, 64, 2048) 0           batch_normalization_97[0][0]     \n",
      "                                                                 batch_normalization_98[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_94 (Activation)      (None, 64, 64, 2048) 0           add_31[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_99 (Conv2D)              (None, 64, 64, 512)  1048576     activation_94[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_99 (BatchNo (None, 64, 64, 512)  2048        conv2d_99[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_95 (Activation)      (None, 64, 64, 512)  0           batch_normalization_99[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_100 (Conv2D)             (None, 64, 64, 512)  2359296     activation_95[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_100 (BatchN (None, 64, 64, 512)  2048        conv2d_100[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_96 (Activation)      (None, 64, 64, 512)  0           batch_normalization_100[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_101 (Conv2D)             (None, 64, 64, 2048) 1048576     activation_96[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_101 (BatchN (None, 64, 64, 2048) 8192        conv2d_101[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "add_32 (Add)                    (None, 64, 64, 2048) 0           batch_normalization_101[0][0]    \n",
      "                                                                 activation_94[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_97 (Activation)      (None, 64, 64, 2048) 0           add_32[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_102 (Conv2D)             (None, 64, 64, 512)  1048576     activation_97[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_102 (BatchN (None, 64, 64, 512)  2048        conv2d_102[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_98 (Activation)      (None, 64, 64, 512)  0           batch_normalization_102[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_103 (Conv2D)             (None, 64, 64, 512)  2359296     activation_98[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_103 (BatchN (None, 64, 64, 512)  2048        conv2d_103[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_99 (Activation)      (None, 64, 64, 512)  0           batch_normalization_103[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_104 (Conv2D)             (None, 64, 64, 2048) 1048576     activation_99[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_104 (BatchN (None, 64, 64, 2048) 8192        conv2d_104[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "add_33 (Add)                    (None, 64, 64, 2048) 0           batch_normalization_104[0][0]    \n",
      "                                                                 activation_97[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_100 (Activation)     (None, 64, 64, 2048) 0           add_33[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_105 (Conv2D)             (None, 64, 64, 512)  9437184     activation_100[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_105 (BatchN (None, 64, 64, 512)  2048        conv2d_105[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_101 (Activation)     (None, 64, 64, 512)  0           batch_normalization_105[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "pam_1 (PAM)                     (None, 64, 64, 512)  1           activation_101[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "cam_1 (CAM)                     (None, 64, 64, 512)  1           activation_101[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_109 (Conv2D)             (None, 64, 64, 512)  2359296     pam_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_111 (Conv2D)             (None, 64, 64, 512)  2359296     cam_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_106 (BatchN (None, 64, 64, 512)  2048        conv2d_109[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_107 (BatchN (None, 64, 64, 512)  2048        conv2d_111[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_103 (Activation)     (None, 64, 64, 512)  0           batch_normalization_106[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "activation_105 (Activation)     (None, 64, 64, 512)  0           batch_normalization_107[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)             (None, 64, 64, 512)  0           activation_103[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)             (None, 64, 64, 512)  0           activation_105[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_110 (Conv2D)             (None, 64, 64, 512)  2359296     dropout_1[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_112 (Conv2D)             (None, 64, 64, 512)  2359296     dropout_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "add_34 (Add)                    (None, 64, 64, 512)  0           conv2d_110[0][0]                 \n",
      "                                                                 conv2d_112[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dropout_3 (Dropout)             (None, 64, 64, 512)  0           add_34[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_113 (Conv2D)             (None, 64, 64, 512)  262656      dropout_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_108 (BatchN (None, 64, 64, 512)  2048        conv2d_113[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_106 (Activation)     (None, 64, 64, 512)  0           batch_normalization_108[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 64, 64, 1024) 0           activation_22[0][0]              \n",
      "                                                                 activation_106[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_114 (Conv2D)             (None, 64, 64, 512)  4719104     concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_109 (BatchN (None, 64, 64, 512)  2048        conv2d_114[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_107 (Activation)     (None, 64, 64, 512)  0           batch_normalization_109[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_115 (Conv2D)             (None, 64, 64, 512)  2359808     activation_107[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_110 (BatchN (None, 64, 64, 512)  2048        conv2d_115[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_108 (Activation)     (None, 64, 64, 512)  0           batch_normalization_110[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d_1 (UpSampling2D)  (None, 128, 128, 512 0           activation_108[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_116 (Conv2D)             (None, 128, 128, 256 524544      up_sampling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_111 (BatchN (None, 128, 128, 256 1024        conv2d_116[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_109 (Activation)     (None, 128, 128, 256 0           batch_normalization_111[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_2 (Concatenate)     (None, 128, 128, 512 0           activation_10[0][0]              \n",
      "                                                                 activation_109[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_117 (Conv2D)             (None, 128, 128, 256 1179904     concatenate_2[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_112 (BatchN (None, 128, 128, 256 1024        conv2d_117[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_110 (Activation)     (None, 128, 128, 256 0           batch_normalization_112[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_118 (Conv2D)             (None, 128, 128, 256 590080      activation_110[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_113 (BatchN (None, 128, 128, 256 1024        conv2d_118[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_111 (Activation)     (None, 128, 128, 256 0           batch_normalization_113[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d_2 (UpSampling2D)  (None, 256, 256, 256 0           activation_111[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_119 (Conv2D)             (None, 256, 256, 64) 65600       up_sampling2d_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_114 (BatchN (None, 256, 256, 64) 256         conv2d_119[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_112 (Activation)     (None, 256, 256, 64) 0           batch_normalization_114[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_3 (Concatenate)     (None, 256, 256, 128 0           activation_1[0][0]               \n",
      "                                                                 activation_112[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_120 (Conv2D)             (None, 256, 256, 64) 73792       concatenate_3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_115 (BatchN (None, 256, 256, 64) 256         conv2d_120[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_113 (Activation)     (None, 256, 256, 64) 0           batch_normalization_115[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_121 (Conv2D)             (None, 256, 256, 64) 36928       activation_113[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_116 (BatchN (None, 256, 256, 64) 256         conv2d_121[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_114 (Activation)     (None, 256, 256, 64) 0           batch_normalization_116[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "up_sampling2d_3 (UpSampling2D)  (None, 512, 512, 64) 0           activation_114[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_122 (Conv2D)             (None, 512, 512, 64) 16448       up_sampling2d_3[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_117 (BatchN (None, 512, 512, 64) 256         conv2d_122[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_115 (Activation)     (None, 512, 512, 64) 0           batch_normalization_117[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_123 (Conv2D)             (None, 512, 512, 64) 36928       activation_115[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_118 (BatchN (None, 512, 512, 64) 256         conv2d_123[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_116 (Activation)     (None, 512, 512, 64) 0           batch_normalization_118[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_124 (Conv2D)             (None, 512, 512, 64) 36928       activation_116[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_119 (BatchN (None, 512, 512, 64) 256         conv2d_124[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "activation_117 (Activation)     (None, 512, 512, 64) 0           batch_normalization_119[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_125 (Conv2D)             (None, 512, 512, 1)  65          activation_117[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_120 (BatchN (None, 512, 512, 1)  4           conv2d_125[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "Classification (Activation)     (None, 512, 512, 1)  0           batch_normalization_120[0][0]    \n",
      "==================================================================================================\n",
      "Total params: 71,399,559\n",
      "Trainable params: 71,285,765\n",
      "Non-trainable params: 113,794\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = danet_resnet101(ImgHeight, ImgWidth, NumChannels, NumClass)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class ParallelModelCheckpoint(ModelCheckpoint):\n",
    "    def __init__(self,model,filepath, monitor='val_loss', verbose=0,\n",
    "                 save_best_only=False, save_weights_only=False,\n",
    "                 mode='auto', period=1):\n",
    "        self.single_model = model\n",
    "        super(ParallelModelCheckpoint,self).__init__(filepath, monitor, verbose,save_best_only, save_weights_only,mode, period)\n",
    "\n",
    "    def set_model(self, model):\n",
    "        super(ParallelModelCheckpoint,self).set_model(self.single_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "paralle_model = multi_gpu_model(model, gpus=4)\n",
    "paralle_model.compile(optimizer=Optimizer, loss=Loss, metrics=Metrics)\n",
    "model_path = \"danet_resnet101-{epoch:02d}-{val_loss:.4f}-{val_acc:.4f}-{val_iou:.4f}.hdf5\"\n",
    "model_checkpoint = ParallelModelCheckpoint(model, model_path, monitor='val_loss', mode='min', verbose=1, save_best_only=False)\n",
    "# model_checkpoint = ModelCheckpoint(model_path, monitor='val_loss', mode='min', verbose=1, save_best_only=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "early_stop = EarlyStopping(monitor='val_loss', mode='min', patience=Patience)\n",
    "reduce_lr = ReduceLROnPlateau(monitor='val_loss', mode='min', factor=0.1, patience=2)\n",
    "check_point_list = [model_checkpoint, early_stop, reduce_lr]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_paths, val_paths = get_data_paths(TrainImageDir, ValImageDir)\n",
    "train_steps = len(train_paths) // BatchSize\n",
    "val_steps = len(val_paths) // BatchSize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "5239/5239 [==============================] - 5406s 1s/step - loss: 1.5943 - acc: 0.8639 - iou: 0.3677 - val_loss: 2.4829 - val_acc: 0.7773 - val_iou: 0.3115\n",
      "\n",
      "Epoch 00001: saving model to danet_resnet101-01-2.4829-0.7773-0.3115.hdf5\n",
      "Epoch 2/100\n",
      "5239/5239 [==============================] - 5308s 1s/step - loss: 1.1217 - acc: 0.9113 - iou: 0.5112 - val_loss: 1.7003 - val_acc: 0.8952 - val_iou: 0.2874\n",
      "\n",
      "Epoch 00002: saving model to danet_resnet101-02-1.7003-0.8952-0.2874.hdf5\n",
      "Epoch 3/100\n",
      "5239/5239 [==============================] - 5323s 1s/step - loss: 0.9739 - acc: 0.9204 - iou: 0.5561 - val_loss: 1.7768 - val_acc: 0.8895 - val_iou: 0.2506\n",
      "\n",
      "Epoch 00003: saving model to danet_resnet101-03-1.7768-0.8895-0.2506.hdf5\n",
      "Epoch 4/100\n",
      "5239/5239 [==============================] - 5319s 1s/step - loss: 0.8804 - acc: 0.9260 - iou: 0.5868 - val_loss: 1.8088 - val_acc: 0.8922 - val_iou: 0.1859\n",
      "\n",
      "Epoch 00004: saving model to danet_resnet101-04-1.8088-0.8922-0.1859.hdf5\n",
      "Epoch 5/100\n",
      "5239/5239 [==============================] - 5326s 1s/step - loss: 0.7599 - acc: 0.9369 - iou: 0.6344 - val_loss: 1.1749 - val_acc: 0.9318 - val_iou: 0.4530\n",
      "\n",
      "Epoch 00005: saving model to danet_resnet101-05-1.1749-0.9318-0.4530.hdf5\n",
      "Epoch 6/100\n",
      "5239/5239 [==============================] - 5325s 1s/step - loss: 0.7296 - acc: 0.9396 - iou: 0.6452 - val_loss: 1.1102 - val_acc: 0.9317 - val_iou: 0.4652\n",
      "\n",
      "Epoch 00006: saving model to danet_resnet101-06-1.1102-0.9317-0.4652.hdf5\n",
      "Epoch 7/100\n",
      "5239/5239 [==============================] - 5316s 1s/step - loss: 0.7153 - acc: 0.9405 - iou: 0.6509 - val_loss: 1.1447 - val_acc: 0.9305 - val_iou: 0.4457\n",
      "\n",
      "Epoch 00007: saving model to danet_resnet101-07-1.1447-0.9305-0.4457.hdf5\n",
      "Epoch 8/100\n",
      "5239/5239 [==============================] - 5321s 1s/step - loss: 0.6980 - acc: 0.9416 - iou: 0.6569 - val_loss: 1.1442 - val_acc: 0.9312 - val_iou: 0.4574\n",
      "\n",
      "Epoch 00008: saving model to danet_resnet101-08-1.1442-0.9312-0.4574.hdf5\n",
      "Epoch 9/100\n",
      "5239/5239 [==============================] - 5329s 1s/step - loss: 0.6819 - acc: 0.9424 - iou: 0.6601 - val_loss: 1.2472 - val_acc: 0.9290 - val_iou: 0.4200\n",
      "\n",
      "Epoch 00009: saving model to danet_resnet101-09-1.2472-0.9290-0.4200.hdf5\n",
      "Epoch 10/100\n",
      "5239/5239 [==============================] - 5361s 1s/step - loss: 0.6756 - acc: 0.9433 - iou: 0.6627 - val_loss: 1.2344 - val_acc: 0.9296 - val_iou: 0.4267\n",
      "\n",
      "Epoch 00010: saving model to danet_resnet101-10-1.2344-0.9296-0.4267.hdf5\n",
      "Epoch 11/100\n",
      "5239/5239 [==============================] - 5323s 1s/step - loss: 0.6699 - acc: 0.9436 - iou: 0.6664 - val_loss: 1.2765 - val_acc: 0.9286 - val_iou: 0.4125\n",
      "\n",
      "Epoch 00011: saving model to danet_resnet101-11-1.2765-0.9286-0.4125.hdf5\n",
      "Epoch 12/100\n",
      "5239/5239 [==============================] - 5320s 1s/step - loss: 0.6720 - acc: 0.9433 - iou: 0.6644 - val_loss: 1.2569 - val_acc: 0.9296 - val_iou: 0.4202\n",
      "\n",
      "Epoch 00012: saving model to danet_resnet101-12-1.2569-0.9296-0.4202.hdf5\n",
      "Epoch 13/100\n",
      "5239/5239 [==============================] - 5326s 1s/step - loss: 0.6811 - acc: 0.9432 - iou: 0.6642 - val_loss: 1.2051 - val_acc: 0.9305 - val_iou: 0.4345\n",
      "\n",
      "Epoch 00013: saving model to danet_resnet101-13-1.2051-0.9305-0.4345.hdf5\n",
      "Epoch 14/100\n",
      "5239/5239 [==============================] - 5332s 1s/step - loss: 0.6731 - acc: 0.9436 - iou: 0.6648 - val_loss: 1.2181 - val_acc: 0.9304 - val_iou: 0.4337\n",
      "\n",
      "Epoch 00014: saving model to danet_resnet101-14-1.2181-0.9304-0.4337.hdf5\n",
      "Epoch 15/100\n",
      "5239/5239 [==============================] - 5330s 1s/step - loss: 0.6754 - acc: 0.9434 - iou: 0.6636 - val_loss: 1.3061 - val_acc: 0.9273 - val_iou: 0.4009\n",
      "\n",
      "Epoch 00015: saving model to danet_resnet101-15-1.3061-0.9273-0.4009.hdf5\n",
      "Epoch 16/100\n",
      "5239/5239 [==============================] - 5344s 1s/step - loss: 0.6828 - acc: 0.9433 - iou: 0.6639 - val_loss: 1.2536 - val_acc: 0.9300 - val_iou: 0.4226\n",
      "\n",
      "Epoch 00016: saving model to danet_resnet101-16-1.2536-0.9300-0.4226.hdf5\n"
     ]
    }
   ],
   "source": [
    "result = paralle_model.fit_generator(\n",
    "    generator=batch_generator(train_paths, BatchSize, flag='train'),\n",
    "    steps_per_epoch=train_steps,\n",
    "    epochs=NumEpochs,\n",
    "    verbose=1,\n",
    "    validation_data=batch_generator(val_paths, BatchSize, flag='test'),\n",
    "    validation_steps=val_steps,\n",
    "    callbacks=check_point_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/data/dh_zhu/anaconda3/lib/python3.6/site-packages/matplotlib/legend.py:326: UserWarning: Unrecognized location \"under right\". Falling back on \"best\"; valid locations are\n",
      "\tbest\n",
      "\tupper right\n",
      "\tupper left\n",
      "\tlower left\n",
      "\tlower right\n",
      "\tright\n",
      "\tcenter left\n",
      "\tcenter right\n",
      "\tlower center\n",
      "\tupper center\n",
      "\tcenter\n",
      "\n",
      "  % (loc, '\\n\\t'.join(self.codes)))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQzHcQwihhmK4hxBCDcVwDyGEGvoH+6YgOrXwkIUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f62a42e75f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAD8CAYAAABw1c+bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3Xl8VNX5+PHPM0v2sAXCkpCwI0vY\njCCgKCqLVrFVrOBai6VWC2rVun2r1upPW7+lrdVWrVr1qwKWxVJFFJeKu+z7FpElBEjCmgSyzvn9\ncSYQQkImZJI7zDzv12ted+beO3eeyfKcc88591wxxqCUUipyuJwOQCmlVNPSxK+UUhFGE79SSkUY\nTfxKKRVhNPErpVSE0cSvlFIRRhO/UkpFGE38SikVYTTxK6VUhPE4HUBNWrdubTp16uR0GEopddpY\nunRpvjGmTSD7hmTi79SpE0uWLHE6DKWUOm2IyLZA99WmHqWUijCa+JVSKsJo4ldKqQgTkm38SqnI\nU1ZWRnZ2NsXFxU6HEtJiYmJITU3F6/We8jE08SulQkJ2djaJiYl06tQJEXE6nJBkjGHv3r1kZ2fT\nuXPnUz6ONvUopUJCcXExSUlJmvRPQkRISkpq8FlRnYlfRDqKyCcisl5E1orI7TXsc76IHBSRFf7H\nQ1W2jRWRjSKSJSL3NShapVRY06Rft2D8jAJp6ikH7jLGLBORRGCpiCw0xqyrtt9nxphLqwXoBp4F\nRgHZwGIRmVfDe5VSSjWROmv8xphdxphl/ucFwHogJcDjDwayjDFbjDGlwAzg8lMNtl6KD8Izg+1S\nKaUCkJCQ4HQITaJebfwi0gkYCHxTw+ahIrJSRN4TkT7+dSnAjir7ZFNLoSEik0VkiYgsycvLq09Y\nNdv0AeRvhM0LG34spZQKIwEnfhFJAGYDdxhjDlXbvAxIN8b0B/4KvF35thoOZWo6vjHmBWNMpjEm\ns02bgKabqNmsSfBYO5jzM/t67s/h8fZ2vVJKBcAYwz333EPfvn3JyMhg5syZAOzatYsRI0YwYMAA\n+vbty2effUZFRQU/+clPju77pz/9yeHo6xbQcE4R8WKT/hvGmDnVt1ctCIwx80XkbyLSGlvD71hl\n11Qgp2Eh12HkA5CzAvZl2dcuL7ToCBc82Kgfq5QKnt/+Zy3rcqrXLxumd4dmPHxZn7p3BObMmcOK\nFStYuXIl+fn5nHXWWYwYMYI333yTMWPG8OCDD1JRUcHhw4dZsWIFO3fuZM2aNQAcOHAgqHE3hkBG\n9QjwErDeGDOtln3a+fdDRAb7j7sXWAx0F5HOIhIFTADmBSv4GiV1hQt/4w/MBb4yOP8BaNWlUT9W\nKRU+Pv/8cyZOnIjb7aZt27acd955LF68mLPOOot//vOfPPLII6xevZrExES6dOnCli1bmDJlCgsW\nLKBZs2ZOh1+nQGr8w4HrgdUissK/7gEgDcAY8xwwHviFiJQDR4AJxhgDlIvIL4H3ATfwsjFmbZC/\nw4nWzrU1fRFwR8Pat6HPDxv9Y5VSwRFozbyx2PR1ohEjRrBo0SLeffddrr/+eu655x5uuOEGVq5c\nyfvvv8+zzz7LW2+9xcsvv9zEEddPnYnfGPM5NbfVV93nGeCZWrbNB+afUnSnavhU6DIS3rkdJkyH\nuFZN+vFKqdPbiBEjeP7557nxxhvZt28fixYt4qmnnmLbtm2kpKTws5/9jKKiIpYtW8Yll1xCVFQU\nV155JV27duUnP/mJ0+HXKTynbEg5E5qlwDtA3gYY9kunI1JKnUZ+9KMf8dVXX9G/f39EhD/84Q+0\na9eOV199laeeegqv10tCQgKvvfYaO3fu5KabbsLn8wHwxBNPOBx93aS2UxonZWZmmqDciOXpgdCm\nF0x8s+HHUko1qvXr19OrVy+nwzgt1PSzEpGlxpjMQN4f3nP1pA2D7V+BvyRWSikV7ok/fSgc2Wcv\n5FJKKQWEfeIfZpfbvnQ2DqWUCiHhnfhbdoaEdra5RymlFBDuiV/E1vq3fQkh2ImtlFJOCO/EDzbx\nH9oJB7Y7HYlSSoWE8E/8aUPtUtv5lVIKiITEn9wbYprDdk38SqngOdnc/Vu3bqVv375NGE39hH/i\nd7lsrX+bdvAqpRSE65QN1aUPg00LoDAXEpKdjkYpVZf37oPdq4N7zHYZcPGTtW6+9957SU9P59Zb\nbwXgkUceQURYtGgR+/fvp6ysjMcee4zLL6/fTQSLi4v5xS9+wZIlS/B4PEybNo2RI0eydu1abrrp\nJkpLS/H5fMyePZsOHTrw4x//mOzsbCoqKvjNb37D1Vdf3aCvXZPISPxp/vH827+C3k1z50el1Oll\nwoQJ3HHHHUcT/1tvvcWCBQu48847adasGfn5+Zx99tmMGzeuXjc8f/bZZwFYvXo1GzZsYPTo0Wza\ntInnnnuO22+/nWuvvZbS0lIqKiqYP38+HTp04N133wXg4MHGuXVsZCT+9v3BE2s7eDXxKxX6TlIz\nbywDBw4kNzeXnJwc8vLyaNmyJe3bt+fOO+9k0aJFuFwudu7cyZ49e2jXrl3Ax/3888+ZMmUKAGec\ncQbp6els2rSJoUOH8vjjj5Odnc0VV1xB9+7dycjI4O677+bee+/l0ksv5dxzz22U7xr+bfwAnijo\neJaO7FFKndT48eOZNWsWM2fOZMKECbzxxhvk5eWxdOlSVqxYQdu2bSkuLq7XMWubCPOaa65h3rx5\nxMbGMmbMGD7++GN69OjB0qVLycjI4P777+fRRx8Nxtc6QWQkfrDNPXvWQHHjnDoppU5/EyZMYMaM\nGcyaNYvx48dz8OBBkpOT8Xq9fPLJJ2zbtq3exxwxYgRvvPEGAJs2bWL79u307NmTLVu20KVLF6ZO\nncq4ceNYtWoVOTk5xMXFcd1113H33XezbNmyYH9FIFKaesB28Bof7PgWuo9yOhqlVAjq06cPBQUF\npKSk0L59e6699louu+wyMjMzGTBgAGeccUa9j3nrrbdyyy23kJGRgcfj4ZVXXiE6OpqZM2fy+uuv\n4/V6adeuHQ899BCLFy/mnnvuweVy4fV6+fvf/94I3zKA+fhFpCPwGtAO8AEvGGP+Um2fa4F7/S8L\ngV8YY1b6t20FCoAKoDyQ+aKDNh9/VaWH4cmOMGwqXPRwcI+tlGownY8/cA2djz+QGn85cJcxZpmI\nJAJLRWShMWZdlX2+B84zxuwXkYuBF4AhVbaPNMbkBxJQo4mKg/YDtJ1fKRXxArnn7i5gl/95gYis\nB1KAdVX2qZpNvwZSgxxncKQPg2+eg7Ji8MY4HY1S6jS3evVqrr/++uPWRUdH88033zgUUWDq1cYv\nIp2AgcDJvtUk4L0qrw3wgYgY4HljzAv1jDF40ofBl0/DzqXQabhjYSilamaMqdcYeadlZGSwYsWK\nJv3MYNwuN+BRPSKSAMwG7jDGHKpln5HYxH9vldXDjTGDgIuB20RkRC3vnSwiS0RkSV5eXsBfoF46\n+luftLlHqZATExPD3r17g5LYwpUxhr179xIT07AWi4Bq/CLixSb9N4wxc2rZpx/wInCxMWZvlUBz\n/MtcEZkLDAYWVX+//0zgBbCdu/X8HoGJawXJfXTCNqVCUGpqKtnZ2TRaxS9MxMTEkJrasNb0OhO/\n2POul4D1xphpteyTBswBrjfGbKqyPh5w+fsG4oHRQONckRCo9KGwcgZUlIM7ckazKhXqvF4vnTt3\ndjqMiBBIU89w4HrgAhFZ4X9cIiK3iMgt/n0eApKAv/m3V47FbAt8LiIrgW+Bd40xC4L9JeolfRiU\nFsLuVY6GoZRSTglkVM/nwEl7W4wxNwM317B+C9D/lKNrDFUnbEsZ5GwsSinlgMiZsqFSs/bQspN2\n8CqlIlbkJX6A9OG2xq+jB5RSESgyE3/aUDi8F/I31b2vUkqFmchM/On+dv5tXzgbh1JKOSAyE3+r\nLpDQVu/Dq5SKSJGZ+EX8N2DXDl6lVOSJzMQPtoP3UDYc2O50JEop1aQiOPEPtUtt7lFKRZjITfzJ\nvSGmuXbwKqUiTuQmfpcbOp5tx/MrpVQEidzED7a5J38TFOpsgEqpyBHhid9/Mxat9SulIkhkJ/72\nA8ATq4lfKRVRIjvxe6IgNVM7eJVSESWyEz/Y6Rt2r4biGu8mqZRSYUcTf9pQMD7Y8a3TkSilVJPQ\nxN9xMLg8eh9epVTE0MQfFQ/t++sVvEqpiFFn4heRjiLyiYisF5G1InJ7DfuIiDwtIlkiskpEBlXZ\ndqOIbPY/bgz2FwiKtKGwcwmUFTsdiVJKNbpAavzlwF3GmF7A2cBtItK72j4XA939j8nA3wFEpBXw\nMDAEGAw8LCItgxR78KQPh4pSyFnmdCRKKdXo6kz8xphdxphl/ucFwHogpdpulwOvGetroIWItAfG\nAAuNMfuMMfuBhcDYoH6DYEg72y51mmalVASoVxu/iHQCBgLfVNuUAuyo8jrbv6629TUde7KILBGR\nJXl5TTyFQlwrO2mbJn6lVAQIOPGLSAIwG7jDGFN90LvU8BZzkvUnrjTmBWNMpjEms02bNoGGFTxp\nQ+2Qzorypv9spZRqQgElfhHxYpP+G8aYOTXskg10rPI6Fcg5yfrQkz4MSgtgz2qnI1FKqUYVyKge\nAV4C1htjptWy2zzgBv/onrOBg8aYXcD7wGgRaenv1B3tXxd60vTGLEqpyOAJYJ/hwPXAahFZ4V/3\nAJAGYIx5DpgPXAJkAYeBm/zb9onI74DF/vc9aozZF7zwg6h5CrRItxdyDb3V6WiUUqrR1Jn4jTGf\nU3NbfdV9DHBbLdteBl4+peiaWvpw2PwBGGNvyK6UUmFIr9ytKn0oHM6H/M1OR6KUUo1GE39VacPs\nUqdpVkqFMU38VSV1hfhkvTGLUiqsaeKvSsQ29+jIHqVUGNPEX136cDi4HQ7sqHtfpZQ6DWnir65y\nPL829yilwpQm/ura9oHoZtrBq5QKW5r4q3O57Wyd2s6vlApTmvhrkjYU8jdCUb7TkSilVNBp4q9J\n+nC71HZ+pVQY0sRfkw4DwROjzT1KqbCkib8mnihIydQOXqVUWNLEX5v0YbB7FZQUOB1J3YoPwjOD\n7VIppeqgib826UPB+OxduULdpg9sZ/TmhU5HopQ6DQQyH39kSh0M4rb34e12odPRHFNSAHkbIXcd\nfPEX2LfFFlAAcybDvCnQ8xIY/5KzcSqlQpYm/tpEJ0D7/o0zsqf4ILw4Cm5eCDHNa96n7IhN8Hkb\nbJLP3QC56+10EpU8MeD2QkWZTf7ighYd4YIHgx+zUipsaOI/mfRh8O0/oLwEPNHBO27Vpple42Bv\nFuStt4m98rH/+2M1eZcXWveAjoPhzBsguTe0OQNadoL1/4FZPwUEfOVw/gPQqkvwYlVKhZ06E7+I\nvAxcCuQaY/rWsP0e4Noqx+sFtPHfdnErUABUAOXGmMxgBd4k0ofBV8/AzmW2zb+hZk2CDe9CRYl9\nPXvS8dvFbZN22z6QcRUkn2GTfKsutmZfk7VzISoeknvBjm9g9Wzo88OGx6qUCluB1PhfAZ4BXqtp\nozHmKeApABG5DLiz2n11RxpjTs9LYI9O2PZlwxP/oRyISwJf2bGavLggpgWMuBs6j4Ck7uCNqd9x\nh0+FS56CnBXw5lXQaXjD4lRKhb1A7rm7SEQ6BXi8icD0hgQUUuJa2SaVbV/CuXfV//0+H2z5BJa8\nDBvfA1MBbTMgdy24Y8BXCpf+qWE19JQz7bLzueCJhX3fnfqxlFIRIWjDOUUkDhgLzK6y2gAfiMhS\nEZkcrM9qUunD7JBOX0Xg7ynaa0fc/HUQvH6F7SAe9kuYutze5SsqAS54ALxxsPbt4MTpjYUu58Om\nBfZm8UopVYtgdu5eBnxRrZlnuDEmR0SSgYUissEYs6imN/sLhskAaWlpQQyrgdKG2Rr7njV2lE9t\njIHtX9t9170NFaX2vSMfhN7jjnUOVzbNJCRDv6vhYHbwYu0xBja9Z0cCJfcK3nGVUmElmIl/AtWa\neYwxOf5lrojMBQYDNSZ+Y8wLwAsAmZmZoVNlrWzb3/ZlzYm/+BCsmmkTfu46O5f/mT+BzJ/WnHwr\nm2bAJv+E5ODF2mOMXW5aoIlfKVWroCR+EWkOnAdcV2VdPOAyxhT4n48GHg3G5zWp5qnQIg2+/xSW\n/PPY2PucFTbZr54FZUW2ULjsacgYb0fZOKFZB2iXYYeLnnOnMzEopUJeIMM5pwPnA61FJBt4GPAC\nGGOe8+/2I+ADY0xRlbe2BeaKSOXnvGmMWRC80JtQ2jDY+K69anbhw3YOn51LbWdqxpWQOQlSBjkd\npdVjLHz2Rzi8z3ZOK6VUNWJCsCMwMzPTLFmyxOkwrFmTYP0822Z/lNga/g3/htgWjoVWo+wl8OKF\ncMWL0O8qp6NRSjUREVka6LVSOklbXUY+AM1SAbGv3VHQujtc9c/QS/oAHQZBXGvbzq+UUjXQxF+X\npK5w0cP2XrzeeHvx1cgHQ3daBJfLdvJmLYSKcqejUUqFIE38gVg71465H3l/cMfeN5buo+1EcNmn\nwZTSSqkmp5O0BaIxx943hq4XgMtjm3vShzkdjVIqxGiNPxApZx4bb5+QHDojeGoT08zeMH7T+05H\nopQKQZr4w1WPsfYK3v1bnY5EKRViNPGHq6NX8X7gbBxKqZCjiT9cJXWFpG46rFMpdYKwTfyHisu4\naNqnHCouczoU5/QYC1s/g5JCpyNRSoWQsE38n2zIJSu3kE825DodinN6jLFXHG/5r9ORKKVCSNgN\n55w6fTkfrN1NaYW9y9Vdb63kvtmrGdW7LU9PHOhwdE0sbaidLXTTAuh1qdPRKKVCRNjV+H81qgfJ\nzWLw+acg8riF1Jax3DW6h7OBOcHthW4X2pu6+3xOR6OUChFhl/g7tY7nvovPAOzsOmXlPu4c1YP0\nJIemSnZa9zFQuBt2r3Q6EqVUiAi7xA/wzqpdxHhcGCDK4+LdVbucDsk53UcBohdzKaWOCsvE//MR\nXVj4q/NoFuPh3O5t+Pl5ITqhWlOIbw2pZ+mwTqXUUWGZ+Pt3bEHHVnFcMSiV/27MI7VlnNMhOavH\nGMhZDgW7nY5EKRUCwjLxV5o4OI3SCh9zloX4pGqNrcdYu9y80Nk4lFIhIawTf892iQxKa8Gb324n\nFO801mTa9oFmKdrco5QCAkj8IvKyiOSKyJpatp8vIgdFZIX/8VCVbWNFZKOIZInIfcEMPFATB6ex\nJa+IxVv3O/HxoUHENvd89wmUlzgdjVLKYYHU+F8Bxtaxz2fGmAH+x6MAIuIGngUuBnoDE0Wkd0OC\nPRU/6NeexGgP07/d3tQfHVp6jIWyItj6udORKKUcVmfiN8YsAvadwrEHA1nGmC3GmFJgBnD5KRyn\nQeKiPPxwYArvrt7FgcOldb8hXHUeAZ5Y2KyzdSoV6YLVxj9URFaKyHsi0se/LgXYUWWfbP+6GonI\nZBFZIiJL8vLyghSWNXFwGqXlPuYu3xnU455WvLHQ5TzY+B5Ecn+HUiooiX8ZkG6M6Q/8Fai8Ia3U\nsG+tGccY84IxJtMYk9mmTZsghHVM7w7N6J/anOmR3snbfTQc2Ab5m5yORCnloAYnfmPMIWNMof/5\nfMArIq2xNfyOVXZNBXIa+nmnauLgNDbtKWTZ9gju5D16cxYd3aNUJGtw4heRdiIi/ueD/cfcCywG\nuotIZxGJAiYA8xr6eafqsv4diI9yM/3bHXXvHK6ap0LbDJ2+QakIF8hwzunAV0BPEckWkUkicouI\n3OLfZTywRkRWAk8DE4xVDvwSeB9YD7xljFnbOF+jbvHRHsYNSOGdVTkcPBLJN2cZA9u/hiMRfOaj\nVIQLZFTPRGNMe2OM1xiTaox5yRjznDHmOf/2Z4wxfYwx/Y0xZxtjvqzy3vnGmB7GmK7GmMcb84sE\n4prBaRSX+fj3igju5O0xFkwFZH3kdCRKKYeE9ZW71WWkNqdvSjPe/CaCO3lTBkFckjb3KBXBIirx\nA0w4K40NuwtYmX3Q6VCc4XLb0T1ZC8FX4XQ0SikHRFziv3xAB2K9bqZ/E8FX8vYYY9v4sxc7HYlS\nygERl/gTY7yM69+B/6zKoaA4Qjt5u14ALo8O61QqQkVc4geYMLgjh0srmLfSscsKnBXT3N6IXdv5\nlYpIEZn4B3RswRntEiN74rYeYyF3Hezf5nQkSqkmFpGJX0S4Zkgaa3YeYnWkdvIevTmLTtqmVKSJ\nyMQPcPmAFGK8LqYvjtBaf+tu0KqrNvcoFYEiNvE3j/Xyg4wO/Hv5TopKyp0Oxxk9xsL3i6C0yOlI\nlFJNKGITP8A1QzpSVFrBO6sitJO3x2ioKIEtnzodiVKqCUV04h+U1pLuyQm8GakTt6UNg6hEHdap\nVISJ6MQvIkwcnMbKHQdYl3PI6XCanicKul1gO3gjdQoLpSJQRCd+gCsGpRDlcTEjUjt5e4yFgl2w\ne5XTkSgV+ooPwjOD7fI0FvGJv0VcFJf0bcfcZTs5UhqBc9d0GwWIju5RKhCbPoD8jbB5odORNEjE\nJ36wd+cqKCmPzE7ehDaQcqa28yt1MrMmwWPtYO5k+3rOz+Cxtnb9aUgTPzC4cyu6tIlnxuII7eTt\nMRZ2LoXCXKcjUU0hTJormkzpYUjqCr5yMD67zvigvBh2fAPz77FnzKfRsGhN/Pg7ec9KY+m2/Wzc\nXeB0OE2v8l68p/npqwpQmDRXNLoj+2HRU/DnDPj099AyHcQFnjgQNwy4DpJ7w/LX4c0fw+87wavj\n4Iu/wJ61IT1gIpBbL74sIrkisqaW7deKyCr/40sR6V9l21YRWS0iK0RkSTADD7Yrz0wlyu2KzPl7\n2mVAYgdt7gl3M66D37WxzRRgl79LhtfHw+F9p56owu0M4tAu+OB/4E994ePH7M2LbnoPkvtAVAJc\n8ABExdsa/rVvwa+/h+vfhiE/h6I8WPgQ/H0YTOsFb98Ga+bYn28IkbruRCUiI4BC4DVjTN8atg8D\n1htj9ovIxcAjxpgh/m1bgUxjTH59gsrMzDRLljR9OTFl+nIWbcrjmwcuJMbrbvLPd9R/7oDV/7J/\nxJ4op6M5vRUfhBdHwc0L7UyoTvBVQN4G2xSxY7Fd7vvu5O/xxEJiO2jWwS4T2/uft/c/9y890ce/\nb9W/YM7NcOVLkDG+8b5TY8vPgi//Aitn2GadvlfC8NttxQhsc2jzjpCQbJtFD2bbQqG6gzvhu48h\n60PY8on9exCX7UvrdhF0vdC+z+UO6t+KiCw1xmQGtG8gtyAUkU7AOzUl/mr7tQTWGGNS/K+3chol\n/i+z8rnmxW/409X9+dHA1Cb/fEdtfA+mT7A1l64jnY7m9OZEIjxyALKXQPa3NslnL4VSf7NlXGvo\nOAQ6ngXlpbbZwh1tr9o+/z5I6maH9B7KgYLdVZ7vsu3Y1cUl2QKgKN/WcDG2zVvctlDoeQmMf6lp\nvncw7FwGX/wZ1s2z8Q+8Dob+Elp1bvixK8ohZ5m9x3XWh7bwwEBsS+gy0ib7pf8Myt9KfRK/p0Gf\ndKJJwHtVXhvgAxExwPPGmBeC/HlBdXaXJDolxTH9mx2Rl/g7nweeGNtJpYn/1MyaBBvetQkVbFPK\n27dC+nC49I8Q08L+o7vqeTZZvVbo88HezbCjMskvtrV7sDXLtn2g34+PJfuWnUHEbn/rRttMcd6v\n4dM/wJ519nlNjIHiA7bpoyDHv9x97Lmv3Cb+ox2e/uHQLTra4yb3Ova5ocYY2PJfm/C3/Beim8O5\nv4Iht9gafbC4PdBxsH2MvN82+Wz5BD78Laydc2y/OZNh3pQmKzSDVuMXkZHA34BzjDF7/es6GGNy\nRCQZWAhMMcYsquX9k4HJAGlpaWdu2+bMPPHPffodT763gQ9/dR7dkhMcicExb1wF+Zth6vLQ/YcN\nZWvm2ERffuQkOwnENLOFQGxLiG1R+/PYlvb11i/g/fug9w9tu3L2YpuQwW7vOBhS/cklZRBEJ9b+\n8YE2VwRq7dsw66f2jm4VpfZMoMA/LDqxg73bW7cLbO02rtWpf86pOqHQrID1/4HP/wS7VkBCWxh6\nG5x5k/29NJW938GbE+DANltR8MRCyzSYOANadTmlQzZ5jV9E+gEvAhdXJn0AY0yOf5krInOBwUCN\nid9/NvAC2KaeYMR1KsafmcofP9jIjG+38z+X9nYqDGf0GGOnb9ibBa27Ox3N6WPvd/DhI7B+nk0u\nFSX+ppRS25TSLsM2xRzZbxN29ecHd/qf77e16NqsexsQaNkJxjxuk31SN3DVY3BeypnHnickN7x2\nu3bu8WcQHYfAmMf8bdwfwYZ3YMXrNu4OA6HbhbaNOzUT3N6GfXYgKkcwbZgPvjI74mZvlk2ul/0F\n+k0Ab0zjx1FdUle44EGYPQm88f5mtwdOOenXV4MTv4ikAXOA640xm6qsjwdcxpgC//PRwKMN/bzG\n1johmlG92zJ7WTb3jO1JtCeCOnm7jwHusqN7NPHXrSjftpcvedkm+vMfsFNffL8osKaU6oyxNfrK\nQmDPOlugFOXZpOWOgVbpDaoVBt3wqXDJU7YA6Xe1PYNongqDbrAPXwXkLLeFwHcfwWfT7BDJ6GbQ\neYT/jOBCW5hVakiHZ3mp/fm9fRt8/6n9uQG8fYtdxrSAq16BXuPq3+QWbGvngjfu2N/K2rehzw+b\n5KMDGdUzHTgfaA3sAR4GvADGmOdE5EXgSqCybabcGJMpIl2Auf51HuBNY8zjgQTlVOdupc8253H9\nS9/y9MSBjOvfwbE4HPG3YfaU/CfvOB1J6Co9DF//DT7/M5QdhjNvhPPug8S2jdOUMnvSsc7YK19q\nsuTQKI4csAXjdx9B1sdw0D98ulXXY4VAUT7M+yVc/BSkDz12hlTTo7KQrNyntLDmzxWXHaF0wzxb\n2w4FQf5bCfqonqbmdOL3+Qzn/e8npLaIY/rksx2LwxEf/taeDv96i21nVsf4KmDldPj4cduO3fMH\ncNHD0KZn433mWzfaZpPKWmHXC+HHrzTe5zUlY2yzS2WzUNaHxzqIT8bltZWTyj6Q2JbVHv51uett\nW747ytb8T/dCsw5OjuoJCy6XMOGsNJ56fyPf5xfRuXW80yE1nR5j4fNp8Lez4bZvnBuHHkqMsYlp\n4UOQu9a2k1/5InQa3vifXVMXmsinAAAWAElEQVRTSrgQsU2Krbvbi59y18PrV0Ghf8SQy2vnkrro\nt3aEUGVi98YFNvig+gimJmxKCXU6ZUMtrjozFbdLIm+65tRMe3ViwS69pB9g10p47XJ440ooK4Lx\n/4SbP2qapA+2kKnsgE1IblizUahL7mU7hsF2eGJgzBN2aGq7DNt3EBUf+Iiz4VNhylIYNsUuh09t\ntNBPN5r4a5HcLIaLeiUza0k2peU+p8NpGrMmwROpxyabmvMzOyPhaToDYYMc2G7HVj8/wnbYjn0S\nblsMfa/Qoa6NqbLDc+T9drn27VM/ViQVmvWkif8kJgxOY29RKQvX7XE6lKYx8gHb2eT2T9lgfHZM\nekEOrHgzdOZiCfbcMFWPd+QAfPAb+GumTTrD74CpK+DsX+hUFk1Ba+lNQtv4T2JE9zaktIhlxuLt\n/KBfe6fDaXxJXW3ynz3JzkBYUQI9x8Lu1fD2L+zIku6j7BwmPcZCVJwzcVadXTIYUyJUHm/+PfY6\nhiMHoP8EGPmgvQpVNZ1gX2egaqSJ/yTcLuHqszoybeEmtu89TFqSQ4muKVUfW+yKgttX2Xlg1sy2\nl5lveMe2wZ5xCfQdb4fhNXZt2BiYca0d+VE5Nnv2zTD353YMeLdRVZpg5PjmmBrXi52faN93drQO\nwKqZdthf1wvhR8817vdRykE6nLMOuw4eYfiTH3PT8M58uimPObcOo1lME1xx6JS6xhb7KmDbF7YQ\nWPdvO3Y6pgX0uszWvjud2/ALY8pL7dwze9bA7jWwZ7VdHqlhaltx+Ud5uKpMK2xqeO5/fdxz3/FX\nyrqiIKlzaF0gpVSAdBx/kN386mK+/X4fh4rL+cuEAVw+IMXpkEJDeamd4GrNLDs5WWkhxCfbIXN9\nx0PqWXY6gZNdiVmUb5uS9qyxy91rbLNLZUL2xNibXbTrC20zbEHz6e/t+mBc0BRuF0ipiKWJP4im\nTl/OgjW7Ka2wI3vcAlEeN6N6t+XpiQMdji6ElB2x7eOrZ9kZPitK7JlDnx/ZM4KPH4VRv7NXTx5N\n9GugcPexYyS2h7Z9/Um+L7TrZ/sdqp5BBPuCpnC+QEpFFE38QbQ1v4hJry5m694iKny2iTi9VRyv\n/nQw6UkRdGFXfRQfgo3zYeHDxyf2SiL2bkbtMqok+gyIT6r72MGeEiHYx1PKIXrlbhB1ah3PXaN7\nMnX6cqI8Qmm5j/zCUnIOFGvir01MMzsqJvUseOPHdupZX5m9ErN5B7jmX6c+zUGwR33oKBIVgXQc\nfwDeWbWLWK+be0b3JM7rxmC47qVvePGzLYTiGVPISOoKF/4GMMeuxLzo0cad20YpVSdN/AH4+Ygu\nfHz3+fxsRBc+/fVIXroxk1G92vLYu+uZMn05RSUnmUM90gXzSkylVFBoG/8pMsbw3KdbeOr9DXRL\nTuD56zMjazK3QGkbulJNoj5t/FrjP0Uiwi/O78prPx1CXkEJ4/76OR9GytQO9aHzpSgVcjTxN9A5\n3Vvznynn0Kl1PDe/toRpH2ykwhd6Z1FKKVVJE38QpLaM41+3DOWqM1N5+uMsJr26mAOHS50OSyml\naqSJP0hivG7+ML4fj/+oL19k5XPZM5+zLueQ02EppdQJAkr8IvKyiOSKyJpatouIPC0iWSKySkQG\nVdl2o4hs9j9uDFbgoUhEuHZIOjN/PpSycsMVf/+CucvD6I5JSqmwEGiN/xVg7Em2Xwx09z8mA38H\nEJFW2JuzDwEGAw+LSMtTDfZ0MSitJf+Zcg79Ultw58yVPDJvLWUVEXIzF6VUyAso8RtjFgE1TI14\n1OXAa8b6GmghIu2BMcBCY8w+Y8x+YCEnL0DCRpvEaN64eQiTzunMK19u5Zp/fE3uoWKnw1JKqaC1\n8acAO6q8zvavq219RPC6Xfzm0t78ZcIA1uw8xKV//Zyl2/ZxqLiMi6Z9yqHiMqdDVEpFoGAl/ppu\nQmpOsv7EA4hMFpElIrIkLy8vSGGFhssHpDD3tmHERrm5+vmvefQ/a8nKLeSTDblOh6aUikDBSvzZ\nQNV71KUCOSdZfwJjzAvGmExjTGabNm2CFFboOKNdM3q3b4bPGGYt3QnAr2auoNdvFjB1+nKHo1NK\nRZJgJf55wA3+0T1nAweNMbuA94HRItLS36k72r8uIt079gw6t47H47InQhUGYrwuJp3TydnAlFIR\nJdDhnNOBr4CeIpItIpNE5BYRucW/y3xgC5AF/AO4FcAYsw/4HbDY/3jUvy4iVU7xDDbhC3DgcBkT\n//ENT763gb2FJc4GqJSKCAHNx2+MmVjHdgPcVsu2l4GX6x9aeKqc4nnqhd15+qPNDExrQYu4KJ5f\n9B2vfbWV64emM/ncLiQlRDsdqlIqTOnsnE1s5Y4DdGgRS5vEaPIKSth18Aj9UluQlVvAXz/OYt7K\nHGI8bm4Yms7PRnShtRYASqkA6K0XT2NZuYU88/Fm5q3MIdrjtmcAWgAopeqgiT8MfJdXyDMfZ/Hv\nFTuJ9ri57uw0Jo/oSptELQCUUifSxB9GtvgLgLdX7CTK4+K6IelMPq8LyYkxToemlAoheiOWMNKl\nTQLTrh7Ah786j0sy2vPyF98z4g+f8Lt31pFbYKeA0CuBlVL1EdCoHuW8Lm0SmPbjAUy5oDvPfJzF\nK19u5fWvt3Hd2el0Soo7eiXw5QMiZkYMpdQp0qae09TW/CKuf+kbduw/cnSdWyDK42ZU77Y8PXGg\ng9EppZqaNvVEgE6t4/m/SUNIbxWH2z8jUoWBFnFe7hrdw9nglFIhTRP/aaxT63juvfgMRIRoj/1V\n7jpYzO8XbDja/q+UUtVp4j/NVV4JfPfoniREuenZNpEP1+dy0R8/5a3FOwjFpjyllLO0jf80V9OV\nwAnRHu6bs5pvv9/HsK5JPHFFBulJ8U6HqpRqRDqOX+HzGWYs3sET89dT5vNx50U9mHROZzxuPclT\nKhxp567C5RKuGZLGwl+dx7nd2/DEexv44d++YG3OQadDU0o5TBN/mGvXPIYXrj+Tv107iN0HSxj3\nzBf8fsEGissqnA5NKeUQTfwRQES4JKM9H/3qPK4clMLf//sdY/+8iK++2+t0aEopB2jijyDN47z8\nYXx/3rh5CD4DE//xNffPWcXBIzrVg1KRRBN/BBrerTXv3zGCn4/owszFOxg17VMWrNntdFhKqSYS\n6K0Xx4rIRhHJEpH7atj+JxFZ4X9sEpEDVbZVVNk2L5jBq1MXG+Xm/kt68e/bziEpIZpbXl/KLf+3\nlNxDxTrpm1Jhrs5J2kTEDTwLjAKygcUiMs8Ys65yH2PMnVX2nwJUnSjmiDFmQPBCVsGUkdqceb8c\nzj8+28KfP9zMF9Py+UFGe530TakwFkiNfzCQZYzZYowpBWYAl59k/4nA9GAEp5qG1+3i1vO7cU63\n1hSVlDNj8Q4A7pixgu4Pzufaf3zN/qJSh6NUSgVLINMypwA7qrzOBobUtKOIpAOdgY+rrI4RkSVA\nOfCkMebtU4xVNbKHLu3Ntr1FbNt3mPIKgwHKKgxffLeXgb9bSOuEKLq2SaBbcgLdkxPolpxIt+QE\n2jaLRkRqPe6h4jKu+NuXzLl1GM1ivE33hZRSNQok8df0H13b5b4TgFnGmKqDxNOMMTki0gX4WERW\nG2O+O+FDRCYDkwHS0tICCEsFW6fW8dw1uidTpy8nLspNSVkFj4zrQ2pLO99/Vm4hm3ML+M/KHA4V\nlx99X2K0h65HC4PKgiGRlJaxuF3CJxtytelIqRBS55QNIjIUeMQYM8b/+n4AY8wTNey7HLjNGPNl\nLcd6BXjHGDPrZJ+pUzY459Y3lvHZpjymXtidpz/azIgebXj22kHH7WOMIa+whKw9hWTlFVYpFArJ\nKyg5up9LoPLPy/hfu13CmWkteeAHvWgZF0XL+Cjio9wnPWOoTs8glDpRUOfqEREPsAm4ENgJLAau\nMcasrbZfT+B9oLPxH1REWgKHjTElItIa+Aq4vGrHcE008Tunpknf+qW2CPj9Bw+X+QuDApZu2887\nq3ZxpLSi1lNEAK9bbCEQF0WLOC+t4qNoERdFq3jv0fUt4712XVwUX23J5/45a/jLhAF6BqGUX9An\naRORS4A/A27gZWPM4yLyKLDEGDPPv88jQIwx5r4q7xsGPA/4sB3JfzbGvFTX52niDx/zV+9i6vTl\neN1CabmP/7m0N/1Sm7O/qIx9h0s5cLiUfUVl/mUpBw6Xsf9wqf9RRoWvrooJtEmI5vyebWiTGE3r\nhGjaJEbTJiGa1on2eWK0R/sgVNirT+IP6J67xpj5wPxq6x6q9vqRGt73JZARyGeo8FR5v4DKpqMl\nW/dz0/DOAb3XGMOh4vKjhcLG3QVMW7iJfUWllPsMboFor5sWcV4+3ZRHfmFpjQVFtMd1fKFQrYDY\nsPsQWbmFvLd6Fz/O7FivZqeaaEHSMJH083Pqu+rN1lWj+vmILvx2XB/aJEbzw4Ep7Dp4pO43+YkI\nzWO9NI/1kp4Uz8C0ljSL9R7tfC4t9/G/V/Xnkoz2gJ2K+sCRMvIKSsgrKCG/0C7zCkvI9y937DvM\n8u372VtUSvWT3Xtnr+be2auJ9brp3jaBpPgoWsVH0zohiqSEKJLio49btoqPIsbrPiHuYHdmh3oi\nDHZ8wfz5BTu2UP6u9aGJXzWq/h2P9Q9U1rYbovoZxLurdh1N/C6X0CreJuSe7RJPepzyCh8rdxzg\njpkr2H2omLIKg8clJMZ4GNK5FcXlPvIL7VlGflEppeW+Go+TGO2xhUFCNDv3HyGvoASfv0S5c+YK\n7v7XSjJSmvOzc7sQ43UT7XUR7XET43XZ1x67rHzureF+CaFekNQ3PmMMh0srKCot53BJBYUl5Rwu\nreCPH2xk2fb9R8/a7py5gnv+tYrBnVvx4A96ERflJjbKTVyUh1ivG7er7jOzYP/sTna80nIfR8oq\nKC6r4EhpBUfK7KPY//ywf1lcVsGMb3ewaU/B0e/6q5kruG/2akb1bsvTEwfW9NFBpTdiUaeVhnY+\nV1fZBxHlcVFa7uPpiQOPFiSVjDEUlpSzr6iU/MJS9haWsLfILvMLS48+332wmG37DtfZL3Eybpcc\nLQyOlNokUfVoLoHkxBjO7NSSaLfraEES7XER5XER7fG/9rqqrHMft/6r7/YybeEm7ryoO8O7tabC\nZ6gwBp8P/9LgM4aKo8tj64/ta3jz2+2szzlkXxs77tvlElJaxJKR2pzDJeUUlfgTfKk/wZeUc7is\n4oSzrVMR7XERV1kQRLltweC1hcPG3QXkHrKFsOFYbGkt4xjePcl+D5/9bpXftep39BlDeZXnG3YV\nsP9wKVV/tQJEeVxEuV0cKaugvAG/92iPkNYqnhdvzDzlu+XpHbiUClAgw1frY/7qXUx5c9nRguS3\nl/dlWNckSsp9FJdVUFzmo6T82LKkyuvisoqj+5WU+8gvLOGzzfkcLi3HZ2zSj3K76NgqDgNV3u8/\nVrkvKAm1IVwCSfFRJMZ6iY/yEB/tJj7KQ1y0h4Rom6Tjo9zER1dfZ/ddum0/T8xfj9fjoqzcx9QL\nu9OvYwuOlB6rMR/xFySV6+z68mPPSys4eKSMnANHjkvGLoH4aA9Rbhcul+AWwe0SXC5wixy3rvLh\n8r8uq/CxeU8hJeUVR38X8VEe/6CCGGKjXMT6z9xiKwugqs/9y6rbP9mQyx0zV5y00lEfQe/cVSpc\nNaQPoibvrNpFXJTnaEHy1Xd7ue7s9FM+XuUZSYzXJodpVw+oNTkYfy21pNxHib/wKCn3UeovGLbm\nF/Hkgg3kFZRQVmHwuoU2CdHcPaYn7ZvH+pMdR5OdS44lvurrKxPjJxtyeXDu6qPJ66/XDGpQ8vrH\nZ98f9/PbtKeQ2y/q0aCfXbASa/Xfxe/H92vY8dbsrrXZsrFp4lcRLdh9EI1RkASaHEQEr1vwul0k\nRJ/4r90vtQUet+u4zvH/ubR3g5LNos35xyXqhiavYP786vOzc+J4wf5bqQ9t6lEqhAW7TyPYTVvB\nji+Ygh1bKH9X0DZ+pVQtQj15qVOnbfxKqRoFu2lLnZ701otKKRVhNPErpVSE0cSvlFIRRhO/UkpF\nGE38SikVYTTxK6VUhNHEr5RSESYkL+ASkTxgWxAO1RrID8JxGkMoxwYaX0NpfA0TyvGFamzpxpg2\ngewYkok/WERkSaBXsjW1UI4NNL6G0vgaJpTjC+XYAqVNPUopFWE08SulVIQJ98T/gtMBnEQoxwYa\nX0NpfA0TyvGFcmwBCes2fqWUUicK9xq/UkqpasIy8YvIWBHZKCJZInKf0/FUJSIdReQTEVkvImtF\n5HanY6qJiLhFZLmIvON0LNWJSAsRmSUiG/w/x6FOx1RJRO70/17XiMh0EYkJgZheFpFcEVlTZV0r\nEVkoIpv9y5YhFNtT/t/tKhGZKyKO3TCgpviqbLtbRIyItHYitoYIu8QvIm7gWeBioDcwUUR6OxvV\nccqBu4wxvYCzgdtCLL5KtwPrnQ6iFn8BFhhjzgD6EyJxikgKMBXINMb0BdzABGejAuAVYGy1dfcB\nHxljugMf+V874RVOjG0h0NcY0w/YBNzf1EFV8QonxoeIdARGAdubOqBgCLvEDwwGsowxW4wxpcAM\n4HKHYzrKGLPLGLPM/7wAm7RSnI3qeCKSCvwAeNHpWKoTkWbACOAlAGNMqTHmgLNRHccDxIqIB4gD\nchyOB2PMImBftdWXA6/6n78K/LBJg/KrKTZjzAfGmHL/y6+B1CYP7FgsNf3sAP4E/Bo4LTtJwzHx\npwA7qrzOJsQSayUR6QQMBL5xNpIT/Bn7R+1zOpAadAHygH/6m6JeFJF4p4MCMMbsBP4XWwvcBRw0\nxnzgbFS1amuM2QW2MgIkOxxPbX4KvOd0EFWJyDhgpzFmpdOxnKpwTPxSw7qQK5VFJAGYDdxhjDnk\ndDyVRORSINcYs9TpWGrhAQYBfzfGDASKcK6Z4jj+dvLLgc5AByBeRK5zNqrTl4g8iG0afcPpWCqJ\nSBzwIPCQ07E0RDgm/mygY5XXqYTA6XZVIuLFJv03jDFznI6nmuHAOBHZim0mu0BEXnc2pONkA9nG\nmMqzpFnYgiAUXAR8b4zJM8aUAXOAYQ7HVJs9ItIewL/MdTie44jIjcClwLUmtMacd8UW7Cv9/yOp\nwDIRaedoVPUUjol/MdBdRDqLSBS2c22ewzEdJSKCbZ9eb4yZ5nQ81Rlj7jfGpBpjOmF/dh8bY0Km\n1mqM2Q3sEJGe/lUXAuscDKmq7cDZIhLn/z1fSIh0PNdgHnCj//mNwL8djOU4IjIWuBcYZ4w57HQ8\nVRljVhtjko0xnfz/I9nAIP/f5Wkj7BK/v1Pol8D72H+6t4wxa52N6jjDgeuxNekV/sclTgd1mpkC\nvCEiq4ABwP9zOB4A/Gchs4BlwGrs/5fjV3mKyHTgK6CniGSLyCTgSWCUiGzGjk55MoRiewZIBBb6\n/z+ecyK2k8R32tMrd5VSKsKEXY1fKaXUyWniV0qpCKOJXymlIowmfqWUijCa+JVSKsJo4ldKqQij\niV8ppSKMJn6llIow/x/ypET3GWVoxgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f62a48807b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure()\n",
    "for metric in metric_list:\n",
    "    plt.plot(result.epoch, result.history[metric], label=metric)\n",
    "    plt.scatter(result.epoch, result.history[metric], marker='*')\n",
    "    val_metric = 'val_' + metric\n",
    "    plt.plot(result.epoch, result.history[val_metric], label=val_metric)\n",
    "    plt.scatter(result.epoch, result.history[val_metric], marker='*')\n",
    "plt.legend(loc='under right')\n",
    "plt.show()\n",
    "\n",
    "plt.figure()\n",
    "plt.plot(result.epoch, result.history['loss'], label=\"loss\")\n",
    "plt.plot(result.epoch, result.history['val_loss'], label=\"val_loss\")\n",
    "plt.scatter(result.epoch, result.history['loss'], marker='*')\n",
    "plt.scatter(result.epoch, result.history['val_loss'], marker='*')\n",
    "plt.legend(loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}