深度残差网络+自适应参数化ReLU激活函数（调参记录4）
续上一篇：
深度残差网络+自适应参数化ReLU激活函数（调参记录3）
https://www.cnblogs.com/shisuzanian/p/12907095.html
本文在深度残差网络中采用了自适应参数化ReLU激活函数，继续测试其在Cifar10图像集上的效果。与上一篇不同的是，这次修改了残差模块里面的结构，原先是两个3×3的卷积层，现在改成了1×1→3×3→1×1的瓶颈式结构，从而网络层数是加深了，但是参数规模减小了。
其中，自适应参数化ReLU是Parametric ReLU的改进版本：
具体Keras代码如下：
1 #!/usr/bin/env python3 2 # -*- coding: utf-8 -*- 3 """ 4 Created on Tue Apr 14 04:17:45 2020 5 Implemented using TensorFlow 1.10.0 and Keras 2.2.1 6 7 Minghang Zhao, Shisheng Zhong, Xuyun Fu, Baoping Tang, Shaojiang Dong, Michael Pecht, 8 Deep Residual Networks with Adaptively Parametric Rectifier Linear Units for Fault Diagnosis, 9 IEEE Transactions on Industrial Electronics, 2020, DOI: 10.1109/TIE.2020.2972458 10 11 @author: Minghang Zhao 12 """ 13 14 from __future__ import print_function 15 import keras 16 import numpy as np 17 from keras.datasets import cifar10 18 from keras.layers import Dense, Conv2D, BatchNormalization, Activation, Minimum 19 from keras.layers import AveragePooling2D, Input, GlobalAveragePooling2D, Concatenate, Reshape 20 from keras.regularizers import l2 21 from keras import backend as K 22 from keras.models import Model 23 from keras import optimizers 24 from keras.preprocessing.image import ImageDataGenerator 25 from keras.callbacks import LearningRateScheduler 26 K.set_learning_phase(1) 27 28 # The data, split between train and test sets 29 (x_train, y_train), (x_test, y_test) = cifar10.load_data() 30 31 # Noised data 32 x_train = x_train.astype('float32') / 255. 33 x_test = x_test.astype('float32') / 255. 34 x_test = x_test-np.mean(x_train) 35 x_train = x_train-np.mean(x_train) 36 print('x_train shape:', x_train.shape) 37 print(x_train.shape[0], 'train samples') 38 print(x_test.shape[0], 'test samples') 39 40 # convert class vectors to binary class matrices 41 y_train = keras.utils.to_categorical(y_train, 10) 42 y_test = keras.utils.to_categorical(y_test, 10) 43 44 # Schedule the learning rate, multiply 0.1 every 200 epoches 45 def scheduler(epoch): 46 if epoch % 200 == 0 and epoch != 0: 47 lr = K.get_value(model.optimizer.lr) 48 K.set_value(model.optimizer.lr, lr * 0.1) 49 print("lr changed to {}".format(lr * 0.1)) 50 return K.get_value(model.optimizer.lr) 51 52 # An adaptively parametric rectifier linear unit (APReLU) 53 def aprelu(inputs): 54 # get the number of channels 55 channels = inputs.get_shape().as_list()[-1] 56 # get a zero feature map 57 zeros_input = keras.layers.subtract([inputs, inputs]) 58 # get a feature map with only positive features 59 pos_input = Activation('relu')(inputs) 60 # get a feature map with only negative features 61 neg_input = Minimum()([inputs,zeros_input]) 62 # define a network to obtain the scaling coefficients 63 scales_p = GlobalAveragePooling2D()(pos_input) 64 scales_n = GlobalAveragePooling2D()(neg_input) 65 scales = Concatenate()([scales_n, scales_p]) 66 scales = Dense(channels//4, activation='linear', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(scales) 67 scales = BatchNormalization()(scales) 68 scales = Activation('relu')(scales) 69 scales = Dense(channels, activation='linear', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(scales) 70 scales = BatchNormalization()(scales) 71 scales = Activation('sigmoid')(scales) 72 scales = Reshape((1,1,channels))(scales) 73 # apply a paramtetric relu 74 neg_part = keras.layers.multiply([scales, neg_input]) 75 return keras.layers.add([pos_input, neg_part]) 76 77 # Residual Block 78 def residual_block(incoming, nb_blocks, out_channels, downsample=False, 79 downsample_strides=2): 80 81 residual = incoming 82 in_channels = incoming.get_shape().as_list()[-1] 83 84 for i in range(nb_blocks): 85 86 identity = residual 87 88 if not downsample: 89 downsample_strides = 1 90 91 residual = BatchNormalization()(residual) 92 residual = aprelu(residual) 93 residual = Conv2D(out_channels//4, 1, strides=(downsample_strides, downsample_strides), 94 padding='same', kernel_initializer='he_normal', 95 kernel_regularizer=l2(1e-4))(residual) 96 97 residual = BatchNormalization()(residual) 98 residual = aprelu(residual) 99 residual = Conv2D(out_channels//4, 3, padding='same', kernel_initializer='he_normal', 100 kernel_regularizer=l2(1e-4))(residual) 101 102 residual = BatchNormalization()(residual) 103 residual = aprelu(residual) 104 residual = Conv2D(out_channels, 1, padding='same', kernel_initializer='he_normal', 105 kernel_regularizer=l2(1e-4))(residual) 106 107 # Downsampling 108 if downsample_strides > 1: 109 identity = AveragePooling2D(pool_size=(1,1), strides=(2,2))(identity) 110 111 # Zero_padding to match channels 112 if in_channels != out_channels: 113 zeros_identity = keras.layers.subtract([identity, identity]) 114 identity = keras.layers.concatenate([identity, zeros_identity]) 115 in_channels = out_channels 116 117 residual = keras.layers.add([residual, identity]) 118 119 return residual 120 121 122 # define and train a model 123 inputs = Input(shape=(32, 32, 3)) 124 net = Conv2D(16, 3, padding='same', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(inputs) 125 net = residual_block(net, 9, 16, downsample=False) 126 net = residual_block(net, 1, 32, downsample=True) 127 net = residual_block(net, 8, 32, downsample=False) 128 net = residual_block(net, 1, 64, downsample=True) 129 net = residual_block(net, 8, 64, downsample=False) 130 net = BatchNormalization()(net) 131 net = aprelu(net) 132 net = GlobalAveragePooling2D()(net) 133 outputs = Dense(10, activation='softmax', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(net) 134 model = Model(inputs=inputs, outputs=outputs) 135 sgd = optimizers.SGD(lr=0.1, decay=0., momentum=0.9, nesterov=True) 136 model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy']) 137 138 # data augmentation 139 datagen = ImageDataGenerator( 140 # randomly rotate images in the range (deg 0 to 180) 141 rotation_range=30, 142 # randomly flip images 143 horizontal_flip=True, 144 # randomly shift images horizontally 145 width_shift_range=0.125, 146 # randomly shift images vertically 147 height_shift_range=0.125) 148 149 reduce_lr = LearningRateScheduler(scheduler) 150 # fit the model on the batches generated by datagen.flow(). 151 model.fit_generator(datagen.flow(x_train, y_train, batch_size=100), 152 validation_data=(x_test, y_test), epochs=500, 153 verbose=1, callbacks=[reduce_lr], workers=4) 154 155 # get results 156 K.set_learning_phase(0) 157 DRSN_train_score1 = model.evaluate(x_train, y_train, batch_size=100, verbose=0) 158 print('Train loss:', DRSN_train_score1[0]) 159 print('Train accuracy:', DRSN_train_score1[1]) 160 DRSN_test_score1 = model.evaluate(x_test, y_test, batch_size=100, verbose=0) 161 print('Test loss:', DRSN_test_score1[0]) 162 print('Test accuracy:', DRSN_test_score1[1])
实验结果如下：
1 Using TensorFlow backend. 2 x_train shape: (50000, 32, 32, 3) 3 50000 train samples 4 10000 test samples 5 Epoch 1/500 6 120s 241ms/step - loss: 2.3085 - acc: 0.3898 - val_loss: 1.9532 - val_acc: 0.5094 7 Epoch 2/500 8 77s 154ms/step - loss: 1.8971 - acc: 0.5130 - val_loss: 1.7076 - val_acc: 0.5678 9 Epoch 3/500 10 77s 154ms/step - loss: 1.6755 - acc: 0.5682 - val_loss: 1.5036 - val_acc: 0.6182 11 Epoch 4/500 12 77s 154ms/step - loss: 1.5174 - acc: 0.6061 - val_loss: 1.3494 - val_acc: 0.6591 13 Epoch 5/500 14 77s 154ms/step - loss: 1.4061 - acc: 0.6334 - val_loss: 1.2835 - val_acc: 0.6646 15 Epoch 6/500 16 77s 154ms/step - loss: 1.3085 - acc: 0.6570 - val_loss: 1.1890 - val_acc: 0.6935 17 Epoch 7/500 18 77s 154ms/step - loss: 1.2315 - acc: 0.6730 - val_loss: 1.1236 - val_acc: 0.7082 19 Epoch 8/500 20 77s 154ms/step - loss: 1.1676 - acc: 0.6870 - val_loss: 1.1081 - val_acc: 0.7100 21 Epoch 9/500 22 77s 154ms/step - loss: 1.1105 - acc: 0.7017 - val_loss: 0.9947 - val_acc: 0.7442 23 Epoch 10/500 24 77s 153ms/step - loss: 1.0784 - acc: 0.7076 - val_loss: 1.0079 - val_acc: 0.7378 25 Epoch 11/500 26 77s 154ms/step - loss: 1.0402 - acc: 0.7166 - val_loss: 0.9686 - val_acc: 0.7456 27 Epoch 12/500 28 77s 154ms/step - loss: 1.0044 - acc: 0.7279 - val_loss: 0.9421 - val_acc: 0.7506 29 Epoch 13/500 30 77s 155ms/step - loss: 0.9791 - acc: 0.7356 - val_loss: 0.9316 - val_acc: 0.7550 31 Epoch 14/500 32 77s 154ms/step - loss: 0.9566 - acc: 0.7431 - val_loss: 0.9106 - val_acc: 0.7567 33 Epoch 15/500 34 77s 154ms/step - loss: 0.9392 - acc: 0.7477 - val_loss: 0.8879 - val_acc: 0.7676 35 Epoch 16/500 36 77s 153ms/step - loss: 0.9217 - acc: 0.7505 - val_loss: 0.8706 - val_acc: 0.7739 37 Epoch 17/500 38 77s 154ms/step - loss: 0.9025 - acc: 0.7599 - val_loss: 0.8551 - val_acc: 0.7766 39 Epoch 18/500 40 77s 153ms/step - loss: 0.8995 - acc: 0.7572 - val_loss: 0.8515 - val_acc: 0.7750 41 Epoch 19/500 42 77s 154ms/step - loss: 0.8803 - acc: 0.7643 - val_loss: 0.8657 - val_acc: 0.7683 43 Epoch 20/500 44 77s 154ms/step - loss: 0.8713 - acc: 0.7682 - val_loss: 0.8249 - val_acc: 0.7861 45 Epoch 21/500 46 77s 154ms/step - loss: 0.8625 - acc: 0.7710 - val_loss: 0.8161 - val_acc: 0.7896 47 Epoch 22/500 48 77s 154ms/step - loss: 0.8532 - acc: 0.7746 - val_loss: 0.8149 - val_acc: 0.7865 49 Epoch 23/500 50 77s 154ms/step - loss: 0.8529 - acc: 0.7745 - val_loss: 0.8192 - val_acc: 0.7913 51 Epoch 24/500 52 77s 153ms/step - loss: 0.8398 - acc: 0.7789 - val_loss: 0.7975 - val_acc: 0.7978 53 Epoch 25/500 54 77s 153ms/step - loss: 0.8343 - acc: 0.7811 - val_loss: 0.8067 - val_acc: 0.7909 55 Epoch 26/500 56 77s 154ms/step - loss: 0.8250 - acc: 0.7831 - val_loss: 0.7864 - val_acc: 0.8016 57 Epoch 27/500 58 77s 154ms/step - loss: 0.8227 - acc: 0.7835 - val_loss: 0.7928 - val_acc: 0.8000 59 Epoch 28/500 60 77s 154ms/step - loss: 0.8189 - acc: 0.7867 - val_loss: 0.7823 - val_acc: 0.8053 61 Epoch 29/500 62 77s 155ms/step - loss: 0.8156 - acc: 0.7869 - val_loss: 0.7825 - val_acc: 0.8014 63 Epoch 30/500 64 77s 154ms/step - loss: 0.8081 - acc: 0.7916 - val_loss: 0.7704 - val_acc: 0.8074 65 Epoch 31/500 66 77s 154ms/step - loss: 0.8014 - acc: 0.7933 - val_loss: 0.7806 - val_acc: 0.8007 67 Epoch 32/500 68 77s 153ms/step - loss: 0.7975 - acc: 0.7931 - val_loss: 0.7764 - val_acc: 0.8056 69 Epoch 33/500 70 77s 154ms/step - loss: 0.7908 - acc: 0.7942 - val_loss: 0.7652 - val_acc: 0.8103 71 Epoch 34/500 72 77s 154ms/step - loss: 0.7939 - acc: 0.7966 - val_loss: 0.7660 - val_acc: 0.8078 73 Epoch 35/500 74 77s 154ms/step - loss: 0.7882 - acc: 0.7990 - val_loss: 0.7669 - val_acc: 0.8069 75 Epoch 36/500 76 77s 155ms/step - loss: 0.7811 - acc: 0.7998 - val_loss: 0.7603 - val_acc: 0.8101 77 Epoch 37/500 78 77s 154ms/step - loss: 0.7745 - acc: 0.8037 - val_loss: 0.7537 - val_acc: 0.8182 79 Epoch 38/500 80 77s 155ms/step - loss: 0.7791 - acc: 0.8000 - val_loss: 0.7441 - val_acc: 0.8194 81 Epoch 39/500 82 77s 153ms/step - loss: 0.7722 - acc: 0.8025 - val_loss: 0.7907 - val_acc: 0.8011 83 Epoch 40/500 84 77s 154ms/step - loss: 0.7683 - acc: 0.8047 - val_loss: 0.7622 - val_acc: 0.8128 85 Epoch 41/500 86 77s 154ms/step - loss: 0.7689 - acc: 0.8057 - val_loss: 0.7767 - val_acc: 0.8015 87 Epoch 42/500 88 77s 154ms/step - loss: 0.7618 - acc: 0.8069 - val_loss: 0.7487 - val_acc: 0.8159 89 Epoch 43/500 90 77s 154ms/step - loss: 0.7587 - acc: 0.8097 - val_loss: 0.7490 - val_acc: 0.8192 91 Epoch 44/500 92 77s 154ms/step - loss: 0.7593 - acc: 0.8096 - val_loss: 0.7403 - val_acc: 0.8170 93 Epoch 45/500 94 77s 154ms/step - loss: 0.7558 - acc: 0.8116 - val_loss: 0.7475 - val_acc: 0.8193 95 Epoch 46/500 96 77s 154ms/step - loss: 0.7565 - acc: 0.8121 - val_loss: 0.7392 - val_acc: 0.8189 97 Epoch 47/500 98 77s 153ms/step - loss: 0.7480 - acc: 0.8127 - val_loss: 0.7472 - val_acc: 0.8176 99 Epoch 48/500 100 77s 154ms/step - loss: 0.7505 - acc: 0.8134 - val_loss: 0.7340 - val_acc: 0.8235 101 Epoch 49/500 102 77s 153ms/step - loss: 0.7404 - acc: 0.8166 - val_loss: 0.7199 - val_acc: 0.8267 103 Epoch 50/500 104 77s 155ms/step - loss: 0.7421 - acc: 0.8150 - val_loss: 0.7194 - val_acc: 0.8267 105 Epoch 51/500 106 77s 153ms/step - loss: 0.7408 - acc: 0.8172 - val_loss: 0.7321 - val_acc: 0.8207 107 Epoch 52/500 108 77s 154ms/step - loss: 0.7364 - acc: 0.8177 - val_loss: 0.7517 - val_acc: 0.8151 109 Epoch 53/500 110 77s 154ms/step - loss: 0.7362 - acc: 0.8194 - val_loss: 0.7171 - val_acc: 0.8279 111 Epoch 54/500 112 77s 153ms/step - loss: 0.7341 - acc: 0.8193 - val_loss: 0.7596 - val_acc: 0.8130 113 Epoch 55/500 114 77s 154ms/step - loss: 0.7354 - acc: 0.8193 - val_loss: 0.7331 - val_acc: 0.8215 115 Epoch 56/500 116 77s 153ms/step - loss: 0.7297 - acc: 0.8224 - val_loss: 0.7168 - val_acc: 0.8315 117 Epoch 57/500 118 77s 154ms/step - loss: 0.7287 - acc: 0.8206 - val_loss: 0.7042 - val_acc: 0.8354 119 Epoch 58/500 120 77s 154ms/step - loss: 0.7267 - acc: 0.8237 - val_loss: 0.7507 - val_acc: 0.8162 121 Epoch 59/500 122 77s 154ms/step - loss: 0.7246 - acc: 0.8241 - val_loss: 0.7273 - val_acc: 0.8239 123 Epoch 60/500 124 77s 154ms/step - loss: 0.7220 - acc: 0.8242 - val_loss: 0.7350 - val_acc: 0.8221 125 Epoch 61/500 126 77s 154ms/step - loss: 0.7167 - acc: 0.8258 - val_loss: 0.7064 - val_acc: 0.8318 127 Epoch 62/500 128 77s 154ms/step - loss: 0.7158 - acc: 0.8277 - val_loss: 0.6990 - val_acc: 0.8348 129 Epoch 63/500 130 77s 153ms/step - loss: 0.7177 - acc: 0.8259 - val_loss: 0.6947 - val_acc: 0.8388 131 Epoch 64/500 132 77s 153ms/step - loss: 0.7143 - acc: 0.8265 - val_loss: 0.7235 - val_acc: 0.8283 133 Epoch 65/500 134 77s 154ms/step - loss: 0.7167 - acc: 0.8254 - val_loss: 0.7047 - val_acc: 0.8342 135 Epoch 66/500 136 77s 153ms/step - loss: 0.7151 - acc: 0.8277 - val_loss: 0.6992 - val_acc: 0.8320 137 Epoch 67/500 138 77s 154ms/step - loss: 0.7085 - acc: 0.8278 - val_loss: 0.7052 - val_acc: 0.8334 139 Epoch 68/500 140 77s 154ms/step - loss: 0.7053 - acc: 0.8295 - val_loss: 0.6973 - val_acc: 0.8396 141 Epoch 69/500 142 77s 154ms/step - loss: 0.7057 - acc: 0.8291 - val_loss: 0.7047 - val_acc: 0.8371 143 Epoch 70/500 144 77s 154ms/step - loss: 0.6973 - acc: 0.8343 - val_loss: 0.6958 - val_acc: 0.8375 145 Epoch 71/500 146 77s 154ms/step - loss: 0.7018 - acc: 0.8310 - val_loss: 0.6887 - val_acc: 0.8405 147 Epoch 72/500 148 77s 154ms/step - loss: 0.7030 - acc: 0.8333 - val_loss: 0.7100 - val_acc: 0.8301 149 Epoch 73/500 150 77s 154ms/step - loss: 0.6993 - acc: 0.8326 - val_loss: 0.7093 - val_acc: 0.8332 151 Epoch 74/500 152 77s 154ms/step - loss: 0.6995 - acc: 0.8319 - val_loss: 0.6969 - val_acc: 0.8350 153 Epoch 75/500 154 77s 154ms/step - loss: 0.6941 - acc: 0.8346 - val_loss: 0.6762 - val_acc: 0.8436 155 Epoch 76/500 156 77s 154ms/step - loss: 0.6976 - acc: 0.8329 - val_loss: 0.7143 - val_acc: 0.8304 157 Epoch 77/500 158 77s 154ms/step - loss: 0.6965 - acc: 0.8335 - val_loss: 0.6836 - val_acc: 0.8411 159 Epoch 78/500 160 77s 154ms/step - loss: 0.6950 - acc: 0.8327 - val_loss: 0.6773 - val_acc: 0.8439 161 Epoch 79/500 162 77s 154ms/step - loss: 0.6961 - acc: 0.8328 - val_loss: 0.6982 - val_acc: 0.8375 163 Epoch 80/500 164 77s 154ms/step - loss: 0.6882 - acc: 0.8368 - val_loss: 0.6908 - val_acc: 0.8396 165 Epoch 81/500 166 77s 153ms/step - loss: 0.6935 - acc: 0.8363 - val_loss: 0.6779 - val_acc: 0.8439 167 Epoch 82/500 168 77s 153ms/step - loss: 0.6927 - acc: 0.8354 - val_loss: 0.6916 - val_acc: 0.8419 169 Epoch 83/500 170 77s 154ms/step - loss: 0.6884 - acc: 0.8391 - val_loss: 0.6962 - val_acc: 0.8402 171 Epoch 84/500 172 77s 154ms/step - loss: 0.6887 - acc: 0.8379 - val_loss: 0.6850 - val_acc: 0.8401 173 Epoch 85/500 174 77s 154ms/step - loss: 0.6843 - acc: 0.8384 - val_loss: 0.6836 - val_acc: 0.8411 175 Epoch 86/500 176 77s 154ms/step - loss: 0.6855 - acc: 0.8383 - val_loss: 0.6807 - val_acc: 0.8445 177 Epoch 87/500 178 77s 153ms/step - loss: 0.6829 - acc: 0.8387 - val_loss: 0.6820 - val_acc: 0.8401 179 Epoch 88/500 180 77s 153ms/step - loss: 0.6790 - acc: 0.8392 - val_loss: 0.6677 - val_acc: 0.8467 181 Epoch 89/500 182 77s 154ms/step - loss: 0.6774 - acc: 0.8402 - val_loss: 0.6831 - val_acc: 0.8440 183 Epoch 90/500 184 77s 154ms/step - loss: 0.6812 - acc: 0.8382 - val_loss: 0.6896 - val_acc: 0.8386 185 Epoch 91/500 186 77s 153ms/step - loss: 0.6746 - acc: 0.8427 - val_loss: 0.6830 - val_acc: 0.8411 187 Epoch 92/500 188 77s 154ms/step - loss: 0.6778 - acc: 0.8405 - val_loss: 0.6687 - val_acc: 0.8468 189 Epoch 93/500 190 77s 154ms/step - loss: 0.6731 - acc: 0.8431 - val_loss: 0.6864 - val_acc: 0.8394 191 Epoch 94/500 192 77s 154ms/step - loss: 0.6788 - acc: 0.8392 - val_loss: 0.6786 - val_acc: 0.8463 193 Epoch 95/500 194 77s 154ms/step - loss: 0.6753 - acc: 0.8423 - val_loss: 0.6808 - val_acc: 0.8412 195 Epoch 96/500 196 77s 154ms/step - loss: 0.6690 - acc: 0.8429 - val_loss: 0.6927 - val_acc: 0.8391 197 Epoch 97/500 198 77s 154ms/step - loss: 0.6753 - acc: 0.8423 - val_loss: 0.6716 - val_acc: 0.8441 199 Epoch 98/500 200 77s 153ms/step - loss: 0.6699 - acc: 0.8422 - val_loss: 0.6747 - val_acc: 0.8440 201 Epoch 99/500 202 76s 152ms/step - loss: 0.6688 - acc: 0.8433 - val_loss: 0.6736 - val_acc: 0.8437 203 Epoch 100/500 204 76s 152ms/step - loss: 0.6634 - acc: 0.8457 - val_loss: 0.6707 - val_acc: 0.8503 205 Epoch 101/500 206 76s 152ms/step - loss: 0.6740 - acc: 0.8415 - val_loss: 0.6442 - val_acc: 0.8537 207 Epoch 102/500 208 76s 152ms/step - loss: 0.6675 - acc: 0.8446 - val_loss: 0.6883 - val_acc: 0.8409 209 Epoch 103/500 210 76s 152ms/step - loss: 0.6691 - acc: 0.8440 - val_loss: 0.6699 - val_acc: 0.8462 211 Epoch 104/500 212 76s 152ms/step - loss: 0.6693 - acc: 0.8440 - val_loss: 0.6707 - val_acc: 0.8458 213 Epoch 105/500 214 76s 152ms/step - loss: 0.6675 - acc: 0.8449 - val_loss: 0.6566 - val_acc: 0.8498 215 Epoch 106/500 216 76s 152ms/step - loss: 0.6672 - acc: 0.8451 - val_loss: 0.6699 - val_acc: 0.8458 217 Epoch 107/500 218 76s 152ms/step - loss: 0.6633 - acc: 0.8457 - val_loss: 0.6869 - val_acc: 0.8418 219 Epoch 108/500 220 76s 153ms/step - loss: 0.6596 - acc: 0.8488 - val_loss: 0.6673 - val_acc: 0.8478 221 Epoch 109/500 222 76s 152ms/step - loss: 0.6624 - acc: 0.8461 - val_loss: 0.6827 - val_acc: 0.8412 223 Epoch 110/500 224 76s 152ms/step - loss: 0.6635 - acc: 0.8460 - val_loss: 0.6767 - val_acc: 0.8430 225 Epoch 111/500 226 76s 152ms/step - loss: 0.6697 - acc: 0.8428 - val_loss: 0.6469 - val_acc: 0.8534 227 Epoch 112/500 228 76s 151ms/step - loss: 0.6627 - acc: 0.8462 - val_loss: 0.6411 - val_acc: 0.8577 229 Epoch 113/500 230 76s 152ms/step - loss: 0.6569 - acc: 0.8489 - val_loss: 0.6673 - val_acc: 0.8461 231 Epoch 114/500 232 76s 152ms/step - loss: 0.6587 - acc: 0.8473 - val_loss: 0.6665 - val_acc: 0.8496 233 Epoch 115/500 234 76s 153ms/step - loss: 0.6560 - acc: 0.8479 - val_loss: 0.6657 - val_acc: 0.8488 235 Epoch 116/500 236 76s 152ms/step - loss: 0.6618 - acc: 0.8453 - val_loss: 0.6782 - val_acc: 0.8442 237 Epoch 117/500 238 76s 152ms/step - loss: 0.6562 - acc: 0.8485 - val_loss: 0.6739 - val_acc: 0.8462 239 Epoch 118/500 240 76s 152ms/step - loss: 0.6620 - acc: 0.8462 - val_loss: 0.6819 - val_acc: 0.8442 241 Epoch 119/500 242 76s 152ms/step - loss: 0.6565 - acc: 0.8486 - val_loss: 0.6531 - val_acc: 0.8522 243 Epoch 120/500 244 76s 152ms/step - loss: 0.6540 - acc: 0.8496 - val_loss: 0.6637 - val_acc: 0.8491 245 Epoch 121/500 246 76s 151ms/step - loss: 0.6567 - acc: 0.8478 - val_loss: 0.6507 - val_acc: 0.8541 247 Epoch 122/500 248 11497s 23s/step - loss: 0.6484 - acc: 0.8514 - val_loss: 0.6679 - val_acc: 0.8465 249 Epoch 123/500 250 76s 152ms/step - loss: 0.6552 - acc: 0.8494 - val_loss: 0.6700 - val_acc: 0.8468 251 Epoch 124/500 252 76s 152ms/step - loss: 0.6600 - acc: 0.8483 - val_loss: 0.6685 - val_acc: 0.8459 253 Epoch 125/500 254 77s 153ms/step - loss: 0.6523 - acc: 0.8499 - val_loss: 0.6754 - val_acc: 0.8435 255 Epoch 126/500 256 76s 152ms/step - loss: 0.6493 - acc: 0.8512 - val_loss: 0.6487 - val_acc: 0.8515 257 Epoch 127/500 258 76s 153ms/step - loss: 0.6507 - acc: 0.8513 - val_loss: 0.6703 - val_acc: 0.8469 259 Epoch 128/500 260 77s 153ms/step - loss: 0.6552 - acc: 0.8484 - val_loss: 0.6527 - val_acc: 0.8506 261 Epoch 129/500 262 76s 153ms/step - loss: 0.6500 - acc: 0.8507 - val_loss: 0.6682 - val_acc: 0.8449 263 Epoch 130/500 264 77s 153ms/step - loss: 0.6534 - acc: 0.8480 - val_loss: 0.6600 - val_acc: 0.8496 265 Epoch 131/500 266 77s 154ms/step - loss: 0.6524 - acc: 0.8507 - val_loss: 0.6506 - val_acc: 0.8505 267 Epoch 132/500 268 76s 152ms/step - loss: 0.6489 - acc: 0.8507 - val_loss: 0.6674 - val_acc: 0.8452 269 Epoch 133/500 270 76s 152ms/step - loss: 0.6499 - acc: 0.8493 - val_loss: 0.6742 - val_acc: 0.8425 271 Epoch 134/500 272 76s 153ms/step - loss: 0.6457 - acc: 0.8519 - val_loss: 0.6522 - val_acc: 0.8516 273 Epoch 135/500 274 76s 152ms/step - loss: 0.6458 - acc: 0.8532 - val_loss: 0.6407 - val_acc: 0.8539 275 Epoch 136/500 276 76s 152ms/step - loss: 0.6478 - acc: 0.8512 - val_loss: 0.6575 - val_acc: 0.8492 277 Epoch 137/500 278 76s 151ms/step - loss: 0.6488 - acc: 0.8508 - val_loss: 0.6673 - val_acc: 0.8456 279 Epoch 138/500 280 76s 152ms/step - loss: 0.6476 - acc: 0.8524 - val_loss: 0.6545 - val_acc: 0.8523 281 Epoch 139/500 282 76s 152ms/step - loss: 0.6517 - acc: 0.8507 - val_loss: 0.6555 - val_acc: 0.8491 283 Epoch 140/500 284 76s 152ms/step - loss: 0.6456 - acc: 0.8531 - val_loss: 0.6658 - val_acc: 0.8460 285 Epoch 141/500 286 76s 152ms/step - loss: 0.6374 - acc: 0.8545 - val_loss: 0.6624 - val_acc: 0.8463 287 Epoch 142/500 288 76s 152ms/step - loss: 0.6437 - acc: 0.8536 - val_loss: 0.6469 - val_acc: 0.8533 289 Epoch 143/500 290 76s 152ms/step - loss: 0.6424 - acc: 0.8520 - val_loss: 0.6703 - val_acc: 0.8469 291 Epoch 144/500 292 76s 152ms/step - loss: 0.6451 - acc: 0.8515 - val_loss: 0.6561 - val_acc: 0.8507 293 Epoch 145/500 294 76s 152ms/step - loss: 0.6472 - acc: 0.8526 - val_loss: 0.6473 - val_acc: 0.8531 295 Epoch 146/500 296 76s 153ms/step - loss: 0.6491 - acc: 0.8518 - val_loss: 0.6320 - val_acc: 0.8589 297 Epoch 147/500 298 76s 152ms/step - loss: 0.6441 - acc: 0.8526 - val_loss: 0.6574 - val_acc: 0.8489 299 Epoch 148/500 300 76s 153ms/step - loss: 0.6453 - acc: 0.8537 - val_loss: 0.6722 - val_acc: 0.8487 301 Epoch 149/500 302 76s 153ms/step - loss: 0.6403 - acc: 0.8539 - val_loss: 0.6543 - val_acc: 0.8572 303 Epoch 150/500 304 76s 153ms/step - loss: 0.6441 - acc: 0.8541 - val_loss: 0.6431 - val_acc: 0.8557 305 Epoch 151/500 306 76s 152ms/step - loss: 0.6407 - acc: 0.8538 - val_loss: 0.6498 - val_acc: 0.8531 307 Epoch 152/500 308 76s 153ms/step - loss: 0.6399 - acc: 0.8538 - val_loss: 0.6524 - val_acc: 0.8497 309 Epoch 153/500 310 76s 152ms/step - loss: 0.6410 - acc: 0.8544 - val_loss: 0.6563 - val_acc: 0.8512 311 Epoch 154/500 312 77s 154ms/step - loss: 0.6456 - acc: 0.8519 - val_loss: 0.6538 - val_acc: 0.8516 313 Epoch 155/500 314 76s 152ms/step - loss: 0.6401 - acc: 0.8558 - val_loss: 0.6553 - val_acc: 0.8501 315 Epoch 156/500 316 76s 152ms/step - loss: 0.6405 - acc: 0.8544 - val_loss: 0.6576 - val_acc: 0.8497 317 Epoch 157/500 318 76s 153ms/step - loss: 0.6401 - acc: 0.8543 - val_loss: 0.6637 - val_acc: 0.8479 319 Epoch 158/500 320 76s 152ms/step - loss: 0.6401 - acc: 0.8553 - val_loss: 0.6510 - val_acc: 0.8554 321 Epoch 159/500 322 76s 152ms/step - loss: 0.6423 - acc: 0.8539 - val_loss: 0.6451 - val_acc: 0.8572 323 Epoch 160/500 324 76s 153ms/step - loss: 0.6376 - acc: 0.8538 - val_loss: 0.6690 - val_acc: 0.8443 325 Epoch 161/500 326 76s 152ms/step - loss: 0.6383 - acc: 0.8558 - val_loss: 0.6621 - val_acc: 0.8492 327 Epoch 162/500 328 76s 152ms/step - loss: 0.6416 - acc: 0.8546 - val_loss: 0.6488 - val_acc: 0.8557 329 Epoch 163/500 330 76s 153ms/step - loss: 0.6386 - acc: 0.8549 - val_loss: 0.6317 - val_acc: 0.8617 331 Epoch 164/500 332 76s 152ms/step - loss: 0.6391 - acc: 0.8552 - val_loss: 0.6382 - val_acc: 0.8588 333 Epoch 165/500 334 76s 153ms/step - loss: 0.6403 - acc: 0.8549 - val_loss: 0.6447 - val_acc: 0.8544 335 Epoch 166/500 336 76s 153ms/step - loss: 0.6400 - acc: 0.8573 - val_loss: 0.6600 - val_acc: 0.8483 337 Epoch 167/500 338 76s 152ms/step - loss: 0.6347 - acc: 0.8560 - val_loss: 0.6413 - val_acc: 0.8535 339 Epoch 168/500 340 76s 152ms/step - loss: 0.6368 - acc: 0.8557 - val_loss: 0.6468 - val_acc: 0.8515 341 Epoch 169/500 342 76s 152ms/step - loss: 0.6349 - acc: 0.8563 - val_loss: 0.6686 - val_acc: 0.8480 343 Epoch 170/500 344 76s 152ms/step - loss: 0.6369 - acc: 0.8557 - val_loss: 0.6449 - val_acc: 0.8560 345 Epoch 171/500 346 76s 152ms/step - loss: 0.6362 - acc: 0.8563 - val_loss: 0.6538 - val_acc: 0.8521 347 Epoch 172/500 348 76s 152ms/step - loss: 0.6321 - acc: 0.8593 - val_loss: 0.6543 - val_acc: 0.8522 349 Epoch 173/500 350 76s 152ms/step - loss: 0.6356 - acc: 0.8569 - val_loss: 0.6445 - val_acc: 0.8512 351 Epoch 174/500 352 77s 154ms/step - loss: 0.6325 - acc: 0.8579 - val_loss: 0.6493 - val_acc: 0.8551 353 Epoch 175/500 354 76s 153ms/step - loss: 0.6330 - acc: 0.8563 - val_loss: 0.6438 - val_acc: 0.8572 355 Epoch 176/500 356 76s 152ms/step - loss: 0.6361 - acc: 0.8547 - val_loss: 0.6432 - val_acc: 0.8532 357 Epoch 177/500 358 76s 152ms/step - loss: 0.6322 - acc: 0.8577 - val_loss: 0.6377 - val_acc: 0.8582 359 Epoch 178/500 360 76s 152ms/step - loss: 0.6476 - acc: 0.8526 - val_loss: 0.6434 - val_acc: 0.8561 361 Epoch 179/500 362 76s 152ms/step - loss: 0.6403 - acc: 0.8540 - val_loss: 0.6569 - val_acc: 0.8529 363 Epoch 180/500 364 76s 153ms/step - loss: 0.6362 - acc: 0.8583 - val_loss: 0.6436 - val_acc: 0.8543 365 Epoch 181/500 366 76s 153ms/step - loss: 0.6300 - acc: 0.8584 - val_loss: 0.6335 - val_acc: 0.8593 367 Epoch 182/500 368 76s 152ms/step - loss: 0.6360 - acc: 0.8565 - val_loss: 0.6460 - val_acc: 0.8554 369 Epoch 183/500 370 76s 152ms/step - loss: 0.6344 - acc: 0.8567 - val_loss: 0.6584 - val_acc: 0.8471 371 Epoch 184/500 372 76s 152ms/step - loss: 0.6354 - acc: 0.8553 - val_loss: 0.6409 - val_acc: 0.8561 373 Epoch 185/500 374 76s 153ms/step - loss: 0.6327 - acc: 0.8578 - val_loss: 0.6422 - val_acc: 0.8590 375 Epoch 186/500 376 76s 151ms/step - loss: 0.6338 - acc: 0.8570 - val_loss: 0.6434 - val_acc: 0.8542 377 Epoch 187/500 378 76s 152ms/step - loss: 0.6283 - acc: 0.8595 - val_loss: 0.6485 - val_acc: 0.8521 379 Epoch 188/500 380 76s 152ms/step - loss: 0.6320 - acc: 0.8565 - val_loss: 0.6415 - val_acc: 0.8560 381 Epoch 189/500 382 76s 152ms/step - loss: 0.6330 - acc: 0.8579 - val_loss: 0.6354 - val_acc: 0.8569 383 Epoch 190/500 384 76s 152ms/step - loss: 0.6260 - acc: 0.8586 - val_loss: 0.6583 - val_acc: 0.8527 385 Epoch 191/500 386 76s 153ms/step - loss: 0.6341 - acc: 0.8577 - val_loss: 0.6381 - val_acc: 0.8549 387 Epoch 192/500 388 77s 154ms/step - loss: 0.6313 - acc: 0.8585 - val_loss: 0.6428 - val_acc: 0.8584 389 Epoch 193/500 390 77s 154ms/step - loss: 0.6297 - acc: 0.8596 - val_loss: 0.6445 - val_acc: 0.8595 391 Epoch 194/500 392 77s 153ms/step - loss: 0.6316 - acc: 0.8579 - val_loss: 0.6446 - val_acc: 0.8578 393 Epoch 195/500 394 77s 154ms/step - loss: 0.6313 - acc: 0.8571 - val_loss: 0.6604 - val_acc: 0.8468 395 Epoch 196/500 396 77s 154ms/step - loss: 0.6287 - acc: 0.8586 - val_loss: 0.6461 - val_acc: 0.8552 397 Epoch 197/500 398 77s 154ms/step - loss: 0.6264 - acc: 0.8597 - val_loss: 0.6453 - val_acc: 0.8543 399 Epoch 198/500 400 77s 154ms/step - loss: 0.6274 - acc: 0.8607 - val_loss: 0.6451 - val_acc: 0.8571 401 Epoch 199/500 402 77s 153ms/step - loss: 0.6314 - acc: 0.8591 - val_loss: 0.6473 - val_acc: 0.8520 403 Epoch 200/500 404 77s 154ms/step - loss: 0.6247 - acc: 0.8619 - val_loss: 0.6640 - val_acc: 0.8488 405 Epoch 201/500 406 lr changed to 0.010000000149011612 407 77s 154ms/step - loss: 0.5292 - acc: 0.8930 - val_loss: 0.5489 - val_acc: 0.8836 408 Epoch 202/500 409 77s 154ms/step - loss: 0.4786 - acc: 0.9093 - val_loss: 0.5324 - val_acc: 0.8892 410 Epoch 203/500 411 77s 154ms/step - loss: 0.4603 - acc: 0.9141 - val_loss: 0.5308 - val_acc: 0.8910 412 Epoch 204/500 413 77s 153ms/step - loss: 0.4479 - acc: 0.9178 - val_loss: 0.5217 - val_acc: 0.8902 414 Epoch 205/500 415 77s 154ms/step - loss: 0.4347 - acc: 0.9205 - val_loss: 0.5181 - val_acc: 0.8903 416 Epoch 206/500 417 77s 154ms/step - loss: 0.4242 - acc: 0.9231 - val_loss: 0.5082 - val_acc: 0.8923 418 Epoch 207/500 419 77s 154ms/step - loss: 0.4196 - acc: 0.9232 - val_loss: 0.5086 - val_acc: 0.8921 420 Epoch 208/500 421 77s 154ms/step - loss: 0.4097 - acc: 0.9255 - val_loss: 0.5067 - val_acc: 0.8932 422 Epoch 209/500 423 77s 154ms/step - loss: 0.4044 - acc: 0.9268 - val_loss: 0.5012 - val_acc: 0.8936 424 Epoch 210/500 425 77s 154ms/step - loss: 0.3980 - acc: 0.9289 - val_loss: 0.5063 - val_acc: 0.8919 426 Epoch 211/500 427 77s 154ms/step - loss: 0.3907 - acc: 0.9294 - val_loss: 0.4907 - val_acc: 0.8964 428 Epoch 212/500 429 77s 154ms/step - loss: 0.3868 - acc: 0.9292 - val_loss: 0.4941 - val_acc: 0.8922 430 Epoch 213/500 431 77s 155ms/step - loss: 0.3798 - acc: 0.9311 - val_loss: 0.4935 - val_acc: 0.8914 432 Epoch 214/500 433 77s 154ms/step - loss: 0.3730 - acc: 0.9321 - val_loss: 0.4874 - val_acc: 0.8955 434 Epoch 215/500 435 77s 154ms/step - loss: 0.3713 - acc: 0.9308 - val_loss: 0.4870 - val_acc: 0.8931 436 Epoch 216/500 437 77s 154ms/step - loss: 0.3670 - acc: 0.9323 - val_loss: 0.4930 - val_acc: 0.8910 438 Epoch 217/500 439 76s 153ms/step - loss: 0.3643 - acc: 0.9325 - val_loss: 0.4798 - val_acc: 0.8938 440 Epoch 218/500 441 76s 152ms/step - loss: 0.3580 - acc: 0.9335 - val_loss: 0.4817 - val_acc: 0.8948 442 Epoch 219/500 443 76s 152ms/step - loss: 0.3548 - acc: 0.9329 - val_loss: 0.4749 - val_acc: 0.8918 444 Epoch 220/500 445 76s 152ms/step - loss: 0.3541 - acc: 0.9334 - val_loss: 0.4663 - val_acc: 0.8966 446 Epoch 221/500 447 76s 153ms/step - loss: 0.3440 - acc: 0.9366 - val_loss: 0.4726 - val_acc: 0.8963 448 Epoch 222/500 449 76s 152ms/step - loss: 0.3434 - acc: 0.9353 - val_loss: 0.4717 - val_acc: 0.8951 450 Epoch 223/500 451 76s 152ms/step - loss: 0.3408 - acc: 0.9355 - val_loss: 0.4629 - val_acc: 0.8976 452 Epoch 224/500 453 76s 153ms/step - loss: 0.3405 - acc: 0.9352 - val_loss: 0.4724 - val_acc: 0.8898 454 Epoch 225/500 455 76s 152ms/step - loss: 0.3355 - acc: 0.9357 - val_loss: 0.4643 - val_acc: 0.8930 456 Epoch 226/500 457 77s 154ms/step - loss: 0.3328 - acc: 0.9363 - val_loss: 0.4663 - val_acc: 0.8962 458 Epoch 227/500 459 76s 152ms/step - loss: 0.3282 - acc: 0.9365 - val_loss: 0.4680 - val_acc: 0.8937 460 Epoch 228/500 461 76s 152ms/step - loss: 0.3307 - acc: 0.9350 - val_loss: 0.4550 - val_acc: 0.8949 462 Epoch 229/500 463 76s 152ms/step - loss: 0.3268 - acc: 0.9350 - val_loss: 0.4638 - val_acc: 0.8967 464 Epoch 230/500 465 76s 152ms/step - loss: 0.3253 - acc: 0.9367 - val_loss: 0.4604 - val_acc: 0.8959 466 Epoch 231/500 467 76s 152ms/step - loss: 0.3191 - acc: 0.9365 - val_loss: 0.4690 - val_acc: 0.8917 468 Epoch 232/500 469 76s 152ms/step - loss: 0.3190 - acc: 0.9369 - val_loss: 0.4653 - val_acc: 0.8924 470 Epoch 233/500 471 76s 152ms/step - loss: 0.3194 - acc: 0.9359 - val_loss: 0.4589 - val_acc: 0.8920 472 Epoch 234/500 473 76s 152ms/step - loss: 0.3107 - acc: 0.9400 - val_loss: 0.4572 - val_acc: 0.8944 474 Epoch 235/500 475 76s 152ms/step - loss: 0.3129 - acc: 0.9367 - val_loss: 0.4646 - val_acc: 0.8925 476 Epoch 236/500 477 76s 152ms/step - loss: 0.3084 - acc: 0.9379 - val_loss: 0.4510 - val_acc: 0.8959 478 Epoch 237/500 479 76s 153ms/step - loss: 0.3114 - acc: 0.9375 - val_loss: 0.4528 - val_acc: 0.8972 480 Epoch 238/500 481 76s 153ms/step - loss: 0.3092 - acc: 0.9380 - val_loss: 0.4624 - val_acc: 0.8928 482 Epoch 239/500 483 76s 152ms/step - loss: 0.3098 - acc: 0.9354 - val_loss: 0.4533 - val_acc: 0.8942 484 Epoch 240/500 485 76s 153ms/step - loss: 0.3027 - acc: 0.9383 - val_loss: 0.4513 - val_acc: 0.8928 486 Epoch 241/500 487 76s 152ms/step - loss: 0.3027 - acc: 0.9385 - val_loss: 0.4576 - val_acc: 0.8927 488 Epoch 242/500 489 76s 152ms/step - loss: 0.3029 - acc: 0.9378 - val_loss: 0.4597 - val_acc: 0.8909 490 Epoch 243/500 491 76s 152ms/step - loss: 0.3023 - acc: 0.9384 - val_loss: 0.4514 - val_acc: 0.8957 492 Epoch 244/500 493 76s 153ms/step - loss: 0.3016 - acc: 0.9366 - val_loss: 0.4510 - val_acc: 0.8932 494 Epoch 245/500 495 76s 152ms/step - loss: 0.3007 - acc: 0.9359 - val_loss: 0.4488 - val_acc: 0.8941 496 Epoch 246/500 497 76s 152ms/step - loss: 0.3017 - acc: 0.9364 - val_loss: 0.4535 - val_acc: 0.8915 498 Epoch 247/500 499 76s 152ms/step - loss: 0.2999 - acc: 0.9368 - val_loss: 0.4524 - val_acc: 0.8925 500 Epoch 248/500 501 76s 152ms/step - loss: 0.3007 - acc: 0.9361 - val_loss: 0.4611 - val_acc: 0.8867 502 Epoch 249/500 503 76s 152ms/step - loss: 0.2982 - acc: 0.9368 - val_loss: 0.4545 - val_acc: 0.8949 504 Epoch 250/500 505 76s 152ms/step - loss: 0.2968 - acc: 0.9371 - val_loss: 0.4599 - val_acc: 0.8892 506 Epoch 251/500 507 76s 152ms/step - loss: 0.2930 - acc: 0.9389 - val_loss: 0.4540 - val_acc: 0.8936 508 Epoch 252/500 509 76s 152ms/step - loss: 0.2904 - acc: 0.9384 - val_loss: 0.4589 - val_acc: 0.8920 510 Epoch 253/500 511 76s 153ms/step - loss: 0.2944 - acc: 0.9373 - val_loss: 0.4548 - val_acc: 0.8906 512 Epoch 254/500 513 76s 152ms/step - loss: 0.2883 - acc: 0.9404 - val_loss: 0.4596 - val_acc: 0.8903 514 Epoch 255/500 515 76s 152ms/step - loss: 0.2917 - acc: 0.9381 - val_loss: 0.4641 - val_acc: 0.8871 516 Epoch 256/500 517 76s 152ms/step - loss: 0.2922 - acc: 0.9368 - val_loss: 0.4643 - val_acc: 0.8868 518 Epoch 257/500 519 76s 152ms/step - loss: 0.2935 - acc: 0.9373 - val_loss: 0.4509 - val_acc: 0.8873 520 Epoch 258/500 521 76s 153ms/step - loss: 0.2934 - acc: 0.9365 - val_loss: 0.4501 - val_acc: 0.8901 522 Epoch 259/500 523 76s 152ms/step - loss: 0.2902 - acc: 0.9381 - val_loss: 0.4459 - val_acc: 0.8928 524 Epoch 260/500 525 76s 152ms/step - loss: 0.2892 - acc: 0.9367 - val_loss: 0.4547 - val_acc: 0.8896 526 Epoch 261/500 527 76s 152ms/step - loss: 0.2892 - acc: 0.9372 - val_loss: 0.4596 - val_acc: 0.8899 528 Epoch 262/500 529 76s 152ms/step - loss: 0.2906 - acc: 0.9360 - val_loss: 0.4500 - val_acc: 0.8889 530 Epoch 263/500 531 76s 152ms/step - loss: 0.2867 - acc: 0.9381 - val_loss: 0.4548 - val_acc: 0.8917 532 Epoch 264/500 533 76s 152ms/step - loss: 0.2906 - acc: 0.9366 - val_loss: 0.4553 - val_acc: 0.8876 534 Epoch 265/500 535 76s 152ms/step - loss: 0.2866 - acc: 0.9377 - val_loss: 0.4549 - val_acc: 0.8914 536 Epoch 266/500 537 76s 153ms/step - loss: 0.2869 - acc: 0.9379 - val_loss: 0.4442 - val_acc: 0.8928 538 Epoch 267/500 539 76s 153ms/step - loss: 0.2883 - acc: 0.9370 - val_loss: 0.4505 - val_acc: 0.8899 540 Epoch 268/500 541 76s 152ms/step - loss: 0.2851 - acc: 0.9388 - val_loss: 0.4590 - val_acc: 0.8879 542 Epoch 269/500 543 76s 152ms/step - loss: 0.2882 - acc: 0.9359 - val_loss: 0.4437 - val_acc: 0.8928 544 Epoch 270/500 545 77s 154ms/step - loss: 0.2882 - acc: 0.9365 - val_loss: 0.4573 - val_acc: 0.8856 546 Epoch 271/500 547 77s 153ms/step - loss: 0.2846 - acc: 0.9385 - val_loss: 0.4599 - val_acc: 0.8881 548 Epoch 272/500 549 76s 153ms/step - loss: 0.2821 - acc: 0.9373 - val_loss: 0.4548 - val_acc: 0.8898 550 Epoch 273/500 551 76s 152ms/step - loss: 0.2878 - acc: 0.9355 - val_loss: 0.4541 - val_acc: 0.8883 552 Epoch 274/500 553 76s 152ms/step - loss: 0.2875 - acc: 0.9357 - val_loss: 0.4588 - val_acc: 0.8881 554 Epoch 275/500 555 76s 152ms/step - loss: 0.2852 - acc: 0.9369 - val_loss: 0.4506 - val_acc: 0.8926 556 Epoch 276/500 557 77s 153ms/step - loss: 0.2867 - acc: 0.9356 - val_loss: 0.4445 - val_acc: 0.8914 558 Epoch 277/500 559 77s 154ms/step - loss: 0.2829 - acc: 0.9374 - val_loss: 0.4466 - val_acc: 0.8913 560 Epoch 278/500 561 76s 152ms/step - loss: 0.2851 - acc: 0.9360 - val_loss: 0.4574 - val_acc: 0.8887 562 Epoch 279/500 563 76s 152ms/step - loss: 0.2868 - acc: 0.9360 - val_loss: 0.4484 - val_acc: 0.8887 564 Epoch 280/500 565 76s 152ms/step - loss: 0.2849 - acc: 0.9369 - val_loss: 0.4615 - val_acc: 0.8851 566 Epoch 281/500 567 76s 152ms/step - loss: 0.2815 - acc: 0.9373 - val_loss: 0.4502 - val_acc: 0.8900 568 Epoch 282/500 569 76s 152ms/step - loss: 0.2863 - acc: 0.9362 - val_loss: 0.4540 - val_acc: 0.8888 570 Epoch 283/500 571 77s 153ms/step - loss: 0.2878 - acc: 0.9362 - val_loss: 0.4559 - val_acc: 0.8872 572 Epoch 284/500 573 76s 152ms/step - loss: 0.2779 - acc: 0.9389 - val_loss: 0.4531 - val_acc: 0.8888 574 Epoch 285/500 575 76s 152ms/step - loss: 0.2801 - acc: 0.9374 - val_loss: 0.4413 - val_acc: 0.8918 576 Epoch 286/500 577 76s 152ms/step - loss: 0.2817 - acc: 0.9380 - val_loss: 0.4584 - val_acc: 0.8864 578 Epoch 287/500 579 76s 152ms/step - loss: 0.2809 - acc: 0.9378 - val_loss: 0.4598 - val_acc: 0.8902 580 Epoch 288/500 581 76s 151ms/step - loss: 0.2784 - acc: 0.9391 - val_loss: 0.4477 - val_acc: 0.8907 582 Epoch 289/500 583 76s 152ms/step - loss: 0.2808 - acc: 0.9370 - val_loss: 0.4581 - val_acc: 0.8877 584 Epoch 290/500 585 76s 152ms/step - loss: 0.2813 - acc: 0.9370 - val_loss: 0.4594 - val_acc: 0.8864 586 Epoch 291/500 587 76s 152ms/step - loss: 0.2795 - acc: 0.9381 - val_loss: 0.4391 - val_acc: 0.8905 588 Epoch 292/500 589 76s 153ms/step - loss: 0.2793 - acc: 0.9385 - val_loss: 0.4471 - val_acc: 0.8881 590 Epoch 293/500 591 76s 153ms/step - loss: 0.2812 - acc: 0.9385 - val_loss: 0.4604 - val_acc: 0.8855 592 Epoch 294/500 593 76s 153ms/step - loss: 0.2808 - acc: 0.9379 - val_loss: 0.4525 - val_acc: 0.8867 594 Epoch 295/500 595 76s 152ms/step - loss: 0.2816 - acc: 0.9373 - val_loss: 0.4532 - val_acc: 0.8873 596 Epoch 296/500 597 76s 153ms/step - loss: 0.2771 - acc: 0.9384 - val_loss: 0.4337 - val_acc: 0.8934 598 Epoch 297/500 599 76s 152ms/step - loss: 0.2793 - acc: 0.9375 - val_loss: 0.4478 - val_acc: 0.8876 600 Epoch 298/500 601 76s 152ms/step - loss: 0.2823 - acc: 0.9375 - val_loss: 0.4560 - val_acc: 0.8889 602 Epoch 299/500 603 76s 153ms/step - loss: 0.2803 - acc: 0.9373 - val_loss: 0.4523 - val_acc: 0.8872 604 Epoch 300/500 605 76s 152ms/step - loss: 0.2796 - acc: 0.9380 - val_loss: 0.4439 - val_acc: 0.8888 606 Epoch 301/500 607 76s 153ms/step - loss: 0.2765 - acc: 0.9388 - val_loss: 0.4537 - val_acc: 0.8881 608 Epoch 302/500 609 76s 152ms/step - loss: 0.2759 - acc: 0.9391 - val_loss: 0.4594 - val_acc: 0.8895 610 Epoch 303/500 611 76s 151ms/step - loss: 0.2822 - acc: 0.9362 - val_loss: 0.4455 - val_acc: 0.8922 612 Epoch 304/500 613 76s 152ms/step - loss: 0.2811 - acc: 0.9361 - val_loss: 0.4593 - val_acc: 0.8870 614 Epoch 305/500 615 76s 152ms/step - loss: 0.2761 - acc: 0.9382 - val_loss: 0.4599 - val_acc: 0.8872 616 Epoch 306/500 617 76s 152ms/step - loss: 0.2753 - acc: 0.9392 - val_loss: 0.4532 - val_acc: 0.8913 618 Epoch 307/500 619 76s 152ms/step - loss: 0.2776 - acc: 0.9393 - val_loss: 0.4373 - val_acc: 0.8916 620 Epoch 308/500 621 76s 152ms/step - loss: 0.2750 - acc: 0.9388 - val_loss: 0.4406 - val_acc: 0.8915 622 Epoch 309/500 623 76s 153ms/step - loss: 0.2778 - acc: 0.9380 - val_loss: 0.4662 - val_acc: 0.8832 624 Epoch 310/500 625 76s 152ms/step - loss: 0.2790 - acc: 0.9384 - val_loss: 0.4385 - val_acc: 0.8960 626 Epoch 311/500 627 76s 152ms/step - loss: 0.2772 - acc: 0.9388 - val_loss: 0.4503 - val_acc: 0.8899 628 Epoch 312/500 629 76s 152ms/step - loss: 0.2776 - acc: 0.9388 - val_loss: 0.4423 - val_acc: 0.8938 630 Epoch 313/500 631 76s 152ms/step - loss: 0.2786 - acc: 0.9379 - val_loss: 0.4404 - val_acc: 0.8951 632 Epoch 314/500 633 76s 153ms/step - loss: 0.2767 - acc: 0.9388 - val_loss: 0.4483 - val_acc: 0.8899 634 Epoch 315/500 635 76s 152ms/step - loss: 0.2741 - acc: 0.9412 - val_loss: 0.4484 - val_acc: 0.8885 636 Epoch 316/500 637 76s 152ms/step - loss: 0.2796 - acc: 0.9371 - val_loss: 0.4526 - val_acc: 0.8883 638 Epoch 317/500 639 76s 152ms/step - loss: 0.2751 - acc: 0.9394 - val_loss: 0.4552 - val_acc: 0.8874 640 Epoch 318/500 641 76s 152ms/step - loss: 0.2775 - acc: 0.9387 - val_loss: 0.4464 - val_acc: 0.8905 642 Epoch 319/500 643 76s 152ms/step - loss: 0.2762 - acc: 0.9388 - val_loss: 0.4523 - val_acc: 0.8889 644 Epoch 320/500 645 76s 152ms/step - loss: 0.2757 - acc: 0.9383 - val_loss: 0.4490 - val_acc: 0.8901 646 Epoch 321/500 647 76s 152ms/step - loss: 0.2732 - acc: 0.9385 - val_loss: 0.4538 - val_acc: 0.8853 648 Epoch 322/500 649 76s 153ms/step - loss: 0.2812 - acc: 0.9377 - val_loss: 0.4450 - val_acc: 0.8909 650 Epoch 323/500 651 76s 153ms/step - loss: 0.2740 - acc: 0.9388 - val_loss: 0.4530 - val_acc: 0.8868 652 Epoch 324/500 653 76s 153ms/step - loss: 0.2730 - acc: 0.9391 - val_loss: 0.4544 - val_acc: 0.8882 654 Epoch 325/500 655 77s 153ms/step - loss: 0.2786 - acc: 0.9385 - val_loss: 0.4564 - val_acc: 0.8881 656 Epoch 326/500 657 76s 152ms/step - loss: 0.2793 - acc: 0.9385 - val_loss: 0.4503 - val_acc: 0.8900 658 Epoch 327/500 659 76s 152ms/step - loss: 0.2764 - acc: 0.9384 - val_loss: 0.4602 - val_acc: 0.8867 660 Epoch 328/500 661 76s 152ms/step - loss: 0.2771 - acc: 0.9386 - val_loss: 0.4446 - val_acc: 0.8888 662 Epoch 329/500 663 76s 152ms/step - loss: 0.2764 - acc: 0.9375 - val_loss: 0.4495 - val_acc: 0.8892 664 Epoch 330/500 665 76s 152ms/step - loss: 0.2773 - acc: 0.9389 - val_loss: 0.4532 - val_acc: 0.8876 666 Epoch 331/500 667 76s 152ms/step - loss: 0.2751 - acc: 0.9399 - val_loss: 0.4550 - val_acc: 0.8890 668 Epoch 332/500 669 76s 152ms/step - loss: 0.2720 - acc: 0.9395 - val_loss: 0.4577 - val_acc: 0.8870 670 Epoch 333/500 671 76s 153ms/step - loss: 0.2713 - acc: 0.9412 - val_loss: 0.4565 - val_acc: 0.8884 672 Epoch 334/500 673 76s 152ms/step - loss: 0.2731 - acc: 0.9399 - val_loss: 0.4496 - val_acc: 0.8904 674 Epoch 335/500 675 76s 152ms/step - loss: 0.2695 - acc: 0.9412 - val_loss: 0.4491 - val_acc: 0.8877 676 Epoch 336/500 677 76s 152ms/step - loss: 0.2715 - acc: 0.9403 - val_loss: 0.4476 - val_acc: 0.8909 678 Epoch 337/500 679 76s 152ms/step - loss: 0.2777 - acc: 0.9365 - val_loss: 0.4533 - val_acc: 0.8889 680 Epoch 338/500 681 76s 152ms/step - loss: 0.2727 - acc: 0.9411 - val_loss: 0.4648 - val_acc: 0.8854 682 Epoch 339/500 683 76s 152ms/step - loss: 0.2712 - acc: 0.9411 - val_loss: 0.4701 - val_acc: 0.8873 684 Epoch 340/500 685 76s 152ms/step - loss: 0.2736 - acc: 0.9398 - val_loss: 0.4632 - val_acc: 0.8874 686 Epoch 341/500 687 77s 153ms/step - loss: 0.2749 - acc: 0.9389 - val_loss: 0.4607 - val_acc: 0.8841 688 Epoch 342/500 689 76s 152ms/step - loss: 0.2697 - acc: 0.9409 - val_loss: 0.4659 - val_acc: 0.8851 690 Epoch 343/500 691 76s 152ms/step - loss: 0.2761 - acc: 0.9391 - val_loss: 0.4545 - val_acc: 0.8854 692 Epoch 344/500 693 76s 152ms/step - loss: 0.2709 - acc: 0.9410 - val_loss: 0.4563 - val_acc: 0.8860 694 Epoch 345/500 695 77s 153ms/step - loss: 0.2746 - acc: 0.9391 - val_loss: 0.4578 - val_acc: 0.8874 696 Epoch 346/500 697 76s 153ms/step - loss: 0.2726 - acc: 0.9406 - val_loss: 0.4714 - val_acc: 0.8847 698 Epoch 347/500 699 77s 153ms/step - loss: 0.2713 - acc: 0.9406 - val_loss: 0.4648 - val_acc: 0.8848 700 Epoch 348/500 701 76s 153ms/step - loss: 0.2745 - acc: 0.9401 - val_loss: 0.4541 - val_acc: 0.8875 702 Epoch 349/500 703 76s 152ms/step - loss: 0.2688 - acc: 0.9421 - val_loss: 0.4635 - val_acc: 0.8840 704 Epoch 350/500 705 76s 152ms/step - loss: 0.2736 - acc: 0.9412 - val_loss: 0.4625 - val_acc: 0.8850 706 Epoch 351/500 707 76s 152ms/step - loss: 0.2721 - acc: 0.9406 - val_loss: 0.4726 - val_acc: 0.8818 708 Epoch 352/500 709 76s 152ms/step - loss: 0.2756 - acc: 0.9399 - val_loss: 0.4567 - val_acc: 0.8870 710 Epoch 353/500 711 76s 152ms/step - loss: 0.2715 - acc: 0.9408 - val_loss: 0.4589 - val_acc: 0.8879 712 Epoch 354/500 713 76s 152ms/step - loss: 0.2714 - acc: 0.9402 - val_loss: 0.4720 - val_acc: 0.8838 714 Epoch 355/500 715 76s 152ms/step - loss: 0.2727 - acc: 0.9398 - val_loss: 0.4646 - val_acc: 0.8861 716 Epoch 356/500 717 76s 152ms/step - loss: 0.2726 - acc: 0.9416 - val_loss: 0.4490 - val_acc: 0.8886 718 Epoch 357/500 719 76s 152ms/step - loss: 0.2715 - acc: 0.9413 - val_loss: 0.4559 - val_acc: 0.8879 720 Epoch 358/500 721 76s 152ms/step - loss: 0.2711 - acc: 0.9414 - val_loss: 0.4723 - val_acc: 0.8867 722 Epoch 359/500 723 76s 152ms/step - loss: 0.2719 - acc: 0.9407 - val_loss: 0.4639 - val_acc: 0.8857 724 Epoch 360/500 725 76s 152ms/step - loss: 0.2745 - acc: 0.9398 - val_loss: 0.4669 - val_acc: 0.8851 726 Epoch 361/500 727 76s 152ms/step - loss: 0.2690 - acc: 0.9413 - val_loss: 0.4633 - val_acc: 0.8860 728 Epoch 362/500 729 76s 152ms/step - loss: 0.2701 - acc: 0.9415 - val_loss: 0.4719 - val_acc: 0.8860 730 Epoch 363/500 731 76s 152ms/step - loss: 0.2712 - acc: 0.9421 - val_loss: 0.4661 - val_acc: 0.8850 732 Epoch 364/500 733 76s 152ms/step - loss: 0.2747 - acc: 0.9393 - val_loss: 0.4545 - val_acc: 0.8875 734 Epoch 365/500 735 77s 153ms/step - loss: 0.2734 - acc: 0.9407 - val_loss: 0.4742 - val_acc: 0.8820 736 Epoch 366/500 737 77s 154ms/step - loss: 0.2745 - acc: 0.9391 - val_loss: 0.4537 - val_acc: 0.8912 738 Epoch 367/500 739 76s 152ms/step - loss: 0.2669 - acc: 0.9422 - val_loss: 0.4615 - val_acc: 0.8867 740 Epoch 368/500 741 76s 152ms/step - loss: 0.2719 - acc: 0.9407 - val_loss: 0.4636 - val_acc: 0.8891 742 Epoch 369/500 743 76s 152ms/step - loss: 0.2706 - acc: 0.9408 - val_loss: 0.4668 - val_acc: 0.8848 744 Epoch 370/500 745 76s 152ms/step - loss: 0.2714 - acc: 0.9404 - val_loss: 0.4527 - val_acc: 0.8901 746 Epoch 371/500 747 76s 152ms/step - loss: 0.2696 - acc: 0.9426 - val_loss: 0.4626 - val_acc: 0.8844 748 Epoch 372/500 749 76s 152ms/step - loss: 0.2662 - acc: 0.9430 - val_loss: 0.4587 - val_acc: 0.8889 750 Epoch 373/500 751 76s 152ms/step - loss: 0.2729 - acc: 0.9410 - val_loss: 0.4603 - val_acc: 0.8879 752 Epoch 374/500 753 76s 152ms/step - loss: 0.2692 - acc: 0.9422 - val_loss: 0.4587 - val_acc: 0.8905 754 Epoch 375/500 755 76s 152ms/step - loss: 0.2719 - acc: 0.9419 - val_loss: 0.4760 - val_acc: 0.8864 756 Epoch 376/500 757 76s 152ms/step - loss: 0.2727 - acc: 0.9401 - val_loss: 0.4500 - val_acc: 0.8895 758 Epoch 377/500 759 76s 151ms/step - loss: 0.2681 - acc: 0.9432 - val_loss: 0.4561 - val_acc: 0.8927 760 Epoch 378/500 761 76s 152ms/step - loss: 0.2763 - acc: 0.9396 - val_loss: 0.4599 - val_acc: 0.8863 762 Epoch 379/500 763 76s 152ms/step - loss: 0.2682 - acc: 0.9413 - val_loss: 0.4728 - val_acc: 0.8849 764 Epoch 380/500 765 76s 152ms/step - loss: 0.2694 - acc: 0.9426 - val_loss: 0.4717 - val_acc: 0.8832 766 Epoch 381/500 767 76s 152ms/step - loss: 0.2710 - acc: 0.9400 - val_loss: 0.4568 - val_acc: 0.8858 768 Epoch 382/500 769 76s 152ms/step - loss: 0.2734 - acc: 0.9393 - val_loss: 0.4745 - val_acc: 0.8831 770 Epoch 383/500 771 76s 152ms/step - loss: 0.2681 - acc: 0.9428 - val_loss: 0.4760 - val_acc: 0.8845 772 Epoch 384/500 773 76s 152ms/step - loss: 0.2720 - acc: 0.9414 - val_loss: 0.4651 - val_acc: 0.8879 774 Epoch 385/500 775 76s 151ms/step - loss: 0.2715 - acc: 0.9412 - val_loss: 0.4527 - val_acc: 0.8924 776 Epoch 386/500 777 76s 152ms/step - loss: 0.2662 - acc: 0.9441 - val_loss: 0.4607 - val_acc: 0.8876 778 Epoch 387/500 779 76s 152ms/step - loss: 0.2649 - acc: 0.9429 - val_loss: 0.4731 - val_acc: 0.8838 780 Epoch 388/500 781 76s 152ms/step - loss: 0.2720 - acc: 0.9407 - val_loss: 0.4683 - val_acc: 0.8842 782 Epoch 389/500 783 76s 152ms/step - loss: 0.2707 - acc: 0.9404 - val_loss: 0.4674 - val_acc: 0.8850 784 Epoch 390/500 785 76s 153ms/step - loss: 0.2687 - acc: 0.9416 - val_loss: 0.4766 - val_acc: 0.8810 786 Epoch 391/500 787 76s 152ms/step - loss: 0.2669 - acc: 0.9440 - val_loss: 0.4728 - val_acc: 0.8834 788 Epoch 392/500 789 77s 153ms/step - loss: 0.2683 - acc: 0.9422 - val_loss: 0.4572 - val_acc: 0.8880 790 Epoch 393/500 791 77s 154ms/step - loss: 0.2631 - acc: 0.9449 - val_loss: 0.4691 - val_acc: 0.8858 792 Epoch 394/500 793 77s 154ms/step - loss: 0.2681 - acc: 0.9419 - val_loss: 0.4747 - val_acc: 0.8875 794 Epoch 395/500 795 77s 154ms/step - loss: 0.2700 - acc: 0.9419 - val_loss: 0.4650 - val_acc: 0.8889 796 Epoch 396/500 797 77s 153ms/step - loss: 0.2702 - acc: 0.9419 - val_loss: 0.4520 - val_acc: 0.8901 798 Epoch 397/500 799 77s 154ms/step - loss: 0.2640 - acc: 0.9439 - val_loss: 0.4607 - val_acc: 0.8857 800 Epoch 398/500 801 77s 154ms/step - loss: 0.2683 - acc: 0.9425 - val_loss: 0.4654 - val_acc: 0.8894 802 Epoch 399/500 803 77s 154ms/step - loss: 0.2709 - acc: 0.9419 - val_loss: 0.4727 - val_acc: 0.8853 804 Epoch 400/500 805 77s 153ms/step - loss: 0.2673 - acc: 0.9429 - val_loss: 0.4670 - val_acc: 0.8873 806 Epoch 401/500 807 lr changed to 0.0009999999776482583 808 77s 154ms/step - loss: 0.2343 - acc: 0.9556 - val_loss: 0.4340 - val_acc: 0.8968 809 Epoch 402/500 810 77s 154ms/step - loss: 0.2155 - acc: 0.9635 - val_loss: 0.4307 - val_acc: 0.9001 811 Epoch 403/500 812 77s 154ms/step - loss: 0.2098 - acc: 0.9645 - val_loss: 0.4287 - val_acc: 0.8996 813 Epoch 404/500 814 77s 153ms/step - loss: 0.2014 - acc: 0.9686 - val_loss: 0.4280 - val_acc: 0.9001 815 Epoch 405/500 816 77s 154ms/step - loss: 0.1992 - acc: 0.9681 - val_loss: 0.4285 - val_acc: 0.9006 817 Epoch 406/500 818 77s 154ms/step - loss: 0.1960 - acc: 0.9695 - val_loss: 0.4308 - val_acc: 0.9000 819 Epoch 407/500 820 77s 153ms/step - loss: 0.1946 - acc: 0.9697 - val_loss: 0.4326 - val_acc: 0.9011 821 Epoch 408/500 822 77s 154ms/step - loss: 0.1956 - acc: 0.9703 - val_loss: 0.4329 - val_acc: 0.9021 823 Epoch 409/500 824 76s 153ms/step - loss: 0.1925 - acc: 0.9713 - val_loss: 0.4312 - val_acc: 0.9020 825 Epoch 410/500 826 77s 153ms/step - loss: 0.1875 - acc: 0.9720 - val_loss: 0.4347 - val_acc: 0.9021 827 Epoch 411/500 828 77s 154ms/step - loss: 0.1895 - acc: 0.9718 - val_loss: 0.4368 - val_acc: 0.9000 829 Epoch 412/500 830 77s 154ms/step - loss: 0.1856 - acc: 0.9722 - val_loss: 0.4390 - val_acc: 0.9012 831 Epoch 413/500 832 77s 154ms/step - loss: 0.1857 - acc: 0.9721 - val_loss: 0.4396 - val_acc: 0.9007 833 Epoch 414/500 834 77s 154ms/step - loss: 0.1842 - acc: 0.9730 - val_loss: 0.4406 - val_acc: 0.9002 835 Epoch 415/500 836 77s 154ms/step - loss: 0.1840 - acc: 0.9734 - val_loss: 0.4426 - val_acc: 0.9003 837 Epoch 416/500 838 77s 154ms/step - loss: 0.1822 - acc: 0.9738 - val_loss: 0.4447 - val_acc: 0.9009 839 Epoch 417/500 840 77s 153ms/step - loss: 0.1828 - acc: 0.9732 - val_loss: 0.4433 - val_acc: 0.8994 841 Epoch 418/500 842 77s 154ms/step - loss: 0.1826 - acc: 0.9735 - val_loss: 0.4407 - val_acc: 0.9006 843 Epoch 419/500 844 77s 153ms/step - loss: 0.1798 - acc: 0.9737 - val_loss: 0.4432 - val_acc: 0.9009 845 Epoch 420/500 846 77s 154ms/step - loss: 0.1800 - acc: 0.9738 - val_loss: 0.4415 - val_acc: 0.9016 847 Epoch 421/500 848 77s 154ms/step - loss: 0.1785 - acc: 0.9743 - val_loss: 0.4447 - val_acc: 0.9012 849 Epoch 422/500 850 77s 154ms/step - loss: 0.1792 - acc: 0.9738 - val_loss: 0.4467 - val_acc: 0.9008 851 Epoch 423/500 852 77s 154ms/step - loss: 0.1763 - acc: 0.9759 - val_loss: 0.4459 - val_acc: 0.9013 853 Epoch 424/500 854 77s 154ms/step - loss: 0.1795 - acc: 0.9735 - val_loss: 0.4501 - val_acc: 0.8997 855 Epoch 425/500 856 76s 153ms/step - loss: 0.1767 - acc: 0.9744 - val_loss: 0.4469 - val_acc: 0.9004 857 Epoch 426/500 858 77s 153ms/step - loss: 0.1766 - acc: 0.9748 - val_loss: 0.4494 - val_acc: 0.9007 859 Epoch 427/500 860 77s 154ms/step - loss: 0.1762 - acc: 0.9748 - val_loss: 0.4534 - val_acc: 0.9001 861 Epoch 428/500 862 77s 153ms/step - loss: 0.1760 - acc: 0.9751 - val_loss: 0.4516 - val_acc: 0.9014 863 Epoch 429/500 864 77s 155ms/step - loss: 0.1752 - acc: 0.9747 - val_loss: 0.4515 - val_acc: 0.8996 865 Epoch 430/500 866 77s 153ms/step - loss: 0.1764 - acc: 0.9747 - val_loss: 0.4529 - val_acc: 0.9010 867 Epoch 431/500 868 77s 154ms/step - loss: 0.1732 - acc: 0.9765 - val_loss: 0.4541 - val_acc: 0.8994 869 Epoch 432/500 870 77s 153ms/step - loss: 0.1720 - acc: 0.9764 - val_loss: 0.4530 - val_acc: 0.9000 871 Epoch 433/500 872 77s 153ms/step - loss: 0.1735 - acc: 0.9756 - val_loss: 0.4527 - val_acc: 0.9007 873 Epoch 434/500 874 77s 154ms/step - loss: 0.1723 - acc: 0.9755 - val_loss: 0.4558 - val_acc: 0.9000 875 Epoch 435/500 876 77s 154ms/step - loss: 0.1731 - acc: 0.9759 - val_loss: 0.4549 - val_acc: 0.9013 877 Epoch 436/500 878 77s 154ms/step - loss: 0.1703 - acc: 0.9764 - val_loss: 0.4560 - val_acc: 0.9017 879 Epoch 437/500 880 77s 155ms/step - loss: 0.1714 - acc: 0.9754 - val_loss: 0.4557 - val_acc: 0.9014 881 Epoch 438/500 882 77s 154ms/step - loss: 0.1691 - acc: 0.9765 - val_loss: 0.4596 - val_acc: 0.8988 883 Epoch 439/500 884 77s 153ms/step - loss: 0.1700 - acc: 0.9761 - val_loss: 0.4613 - val_acc: 0.9006 885 Epoch 440/500 886 77s 154ms/step - loss: 0.1718 - acc: 0.9754 - val_loss: 0.4611 - val_acc: 0.9001 887 Epoch 441/500 888 77s 153ms/step - loss: 0.1704 - acc: 0.9758 - val_loss: 0.4616 - val_acc: 0.9017 889 Epoch 442/500 890 77s 154ms/step - loss: 0.1663 - acc: 0.9781 - val_loss: 0.4638 - val_acc: 0.8990 891 Epoch 443/500 892 77s 154ms/step - loss: 0.1697 - acc: 0.9759 - val_loss: 0.4635 - val_acc: 0.9007 893 Epoch 444/500 894 77s 154ms/step - loss: 0.1673 - acc: 0.9775 - val_loss: 0.4664 - val_acc: 0.8994 895 Epoch 445/500 896 77s 154ms/step - loss: 0.1649 - acc: 0.9779 - val_loss: 0.4651 - val_acc: 0.8991 897 Epoch 446/500 898 77s 153ms/step - loss: 0.1692 - acc: 0.9760 - val_loss: 0.4659 - val_acc: 0.8992 899 Epoch 447/500 900 77s 153ms/step - loss: 0.1678 - acc: 0.9764 - val_loss: 0.4637 - val_acc: 0.8997 901 Epoch 448/500 902 77s 153ms/step - loss: 0.1644 - acc: 0.9774 - val_loss: 0.4659 - val_acc: 0.8996 903 Epoch 449/500 904 77s 153ms/step - loss: 0.1634 - acc: 0.9783 - val_loss: 0.4628 - val_acc: 0.9002 905 Epoch 450/500 906 77s 153ms/step - loss: 0.1662 - acc: 0.9774 - val_loss: 0.4642 - val_acc: 0.9024 907 Epoch 451/500 908 77s 154ms/step - loss: 0.1649 - acc: 0.9767 - val_loss: 0.4647 - val_acc: 0.9020 909 Epoch 452/500 910 77s 153ms/step - loss: 0.1645 - acc: 0.9776 - val_loss: 0.4674 - val_acc: 0.8994 911 Epoch 453/500 912 77s 154ms/step - loss: 0.1646 - acc: 0.9772 - val_loss: 0.4650 - val_acc: 0.8999 913 Epoch 454/500 914 77s 154ms/step - loss: 0.1639 - acc: 0.9786 - val_loss: 0.4683 - val_acc: 0.8973 915 Epoch 455/500 916 77s 154ms/step - loss: 0.1626 - acc: 0.9778 - val_loss: 0.4665 - val_acc: 0.8997 917 Epoch 456/500 918 77s 154ms/step - loss: 0.1634 - acc: 0.9779 - val_loss: 0.4647 - val_acc: 0.8993 919 Epoch 457/500 920 76s 153ms/step - loss: 0.1623 - acc: 0.9785 - val_loss: 0.4645 - val_acc: 0.8996 921 Epoch 458/500 922 77s 154ms/step - loss: 0.1616 - acc: 0.9780 - val_loss: 0.4654 - val_acc: 0.9007 923 Epoch 459/500 924 77s 153ms/step - loss: 0.1617 - acc: 0.9777 - val_loss: 0.4664 - val_acc: 0.8987 925 Epoch 460/500 926 77s 153ms/step - loss: 0.1623 - acc: 0.9777 - val_loss: 0.4652 - val_acc: 0.8989 927 Epoch 461/500 928 77s 154ms/step - loss: 0.1595 - acc: 0.9789 - val_loss: 0.4637 - val_acc: 0.8992 929 Epoch 462/500 930 77s 154ms/step - loss: 0.1609 - acc: 0.9789 - val_loss: 0.4675 - val_acc: 0.8967 931 Epoch 463/500 932 77s 153ms/step - loss: 0.1615 - acc: 0.9779 - val_loss: 0.4731 - val_acc: 0.8981 933 Epoch 464/500 934 77s 153ms/step - loss: 0.1612 - acc: 0.9778 - val_loss: 0.4656 - val_acc: 0.9017 935 Epoch 465/500 936 77s 153ms/step - loss: 0.1571 - acc: 0.9793 - val_loss: 0.4738 - val_acc: 0.9003 937 Epoch 466/500 938 77s 154ms/step - loss: 0.1606 - acc: 0.9773 - val_loss: 0.4741 - val_acc: 0.8996 939 Epoch 467/500 940 76s 153ms/step - loss: 0.1591 - acc: 0.9794 - val_loss: 0.4749 - val_acc: 0.8988 941 Epoch 468/500 942 77s 154ms/step - loss: 0.1594 - acc: 0.9780 - val_loss: 0.4723 - val_acc: 0.8969 943 Epoch 469/500 944 77s 154ms/step - loss: 0.1591 - acc: 0.9786 - val_loss: 0.4748 - val_acc: 0.8981 945 Epoch 470/500 946 77s 154ms/step - loss: 0.1560 - acc: 0.9795 - val_loss: 0.4730 - val_acc: 0.8972 947 Epoch 471/500 948 77s 154ms/step - loss: 0.1574 - acc: 0.9791 - val_loss: 0.4760 - val_acc: 0.8975 949 Epoch 472/500 950 77s 153ms/step - loss: 0.1577 - acc: 0.9786 - val_loss: 0.4757 - val_acc: 0.8974 951 Epoch 473/500 952 77s 153ms/step - loss: 0.1543 - acc: 0.9799 - val_loss: 0.4787 - val_acc: 0.8955 953 Epoch 474/500 954 77s 154ms/step - loss: 0.1552 - acc: 0.9800 - val_loss: 0.4751 - val_acc: 0.8966 955 Epoch 475/500 956 77s 154ms/step - loss: 0.1579 - acc: 0.9778 - val_loss: 0.4761 - val_acc: 0.8954 957 Epoch 476/500 958 77s 154ms/step - loss: 0.1566 - acc: 0.9795 - val_loss: 0.4738 - val_acc: 0.8973 959 Epoch 477/500 960 77s 154ms/step - loss: 0.1552 - acc: 0.9795 - val_loss: 0.4787 - val_acc: 0.8966 961 Epoch 478/500 962 77s 153ms/step - loss: 0.1569 - acc: 0.9789 - val_loss: 0.4724 - val_acc: 0.8986 963 Epoch 479/500 964 77s 154ms/step - loss: 0.1544 - acc: 0.9796 - val_loss: 0.4722 - val_acc: 0.8991 965 Epoch 480/500 966 77s 153ms/step - loss: 0.1566 - acc: 0.9790 - val_loss: 0.4749 - val_acc: 0.8977 967 Epoch 481/500 968 77s 153ms/step - loss: 0.1539 - acc: 0.9797 - val_loss: 0.4756 - val_acc: 0.8982 969 Epoch 482/500 970 77s 154ms/step - loss: 0.1543 - acc: 0.9793 - val_loss: 0.4783 - val_acc: 0.8978 971 Epoch 483/500 972 77s 153ms/step - loss: 0.1546 - acc: 0.9793 - val_loss: 0.4776 - val_acc: 0.8973 973 Epoch 484/500 974 77s 154ms/step - loss: 0.1549 - acc: 0.9787 - val_loss: 0.4755 - val_acc: 0.8977 975 Epoch 485/500 976 77s 154ms/step - loss: 0.1534 - acc: 0.9786 - val_loss: 0.4774 - val_acc: 0.8976 977 Epoch 486/500 978 77s 154ms/step - loss: 0.1528 - acc: 0.9795 - val_loss: 0.4746 - val_acc: 0.8997 979 Epoch 487/500 980 77s 154ms/step - loss: 0.1522 - acc: 0.9798 - val_loss: 0.4762 - val_acc: 0.8996 981 Epoch 488/500 982 77s 153ms/step - loss: 0.1538 - acc: 0.9790 - val_loss: 0.4771 - val_acc: 0.8986 983 Epoch 489/500 984 277/500 [===============>..............] - ETA: 33s - loss: 0.1521 - acc: 0.9798 Traceback (most recent call last): 985 986 File "C:\Users\hitwh\.spyder-py3\temp.py", line 153, in <module> 987 verbose=1, callbacks=[reduce_lr], workers=4) 988 989 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\keras\legacy\interfaces.py", line 91, in wrapper 990 return func(*args, **kwargs) 991 992 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\keras\engine\training.py", line 1415, in fit_generator 993 initial_epoch=initial_epoch) 994 995 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\keras\engine\training_generator.py", line 213, in fit_generator 996 class_weight=class_weight) 997 998 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\keras\engine\training.py", line 1215, in train_on_batch 999 outputs = self.train_function(ins) 1000 1001 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\keras\backend\tensorflow_backend.py", line 2666, in __call__ 1002 return self._call(inputs) 1003 1004 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\keras\backend\tensorflow_backend.py", line 2636, in _call 1005 fetched = self._callable_fn(*array_vals) 1006 1007 File "C:\Users\hitwh\Anaconda3\envs\Initial\lib\site-packages\tensorflow\python\client\session.py", line 1382, in __call__ 1008 run_metadata_ptr) 1009 1010 KeyboardInterrupt
这次是故意中断的，估计跑完500个epoch，效果也没有上一篇（调参记录3）的时候效果好。其中，在第122个epoch的时候，电脑居然休眠了，浪费了一万多秒。
Minghang Zhao, Shisheng Zhong, Xuyun Fu, Baoping Tang, Shaojiang Dong, Michael Pecht, Deep Residual Networks with Adaptively Parametric Rectifier Linear Units for Fault Diagnosis, IEEE Transactions on Industrial Electronics, 2020, DOI: 10.1109/TIE.2020.2972458
