从以往的调参结果来看，过拟合是最主要的问题。本文在调参记录12的基础上，将层数减少，减到9个残差模块，再试一次。
Keras程序如下：
1 #!/usr/bin/env python3 2 # -*- coding: utf-8 -*- 3 """ 4 Created on Tue Apr 14 04:17:45 2020 5 Implemented using TensorFlow 1.0.1 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 1500 epoches 45 def scheduler(epoch): 46 if epoch % 1500 == 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, activation='linear', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(scales) 67 scales = BatchNormalization(momentum=0.9, gamma_regularizer=l2(1e-4))(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(momentum=0.9, gamma_regularizer=l2(1e-4))(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(momentum=0.9, gamma_regularizer=l2(1e-4))(residual) 92 residual = aprelu(residual) 93 residual = Conv2D(out_channels, 3, strides=(downsample_strides, downsample_strides), 94 padding='same', kernel_initializer='he_normal', 95 kernel_regularizer=l2(1e-4))(residual) 96 97 residual = BatchNormalization(momentum=0.9, gamma_regularizer=l2(1e-4))(residual) 98 residual = aprelu(residual) 99 residual = Conv2D(out_channels, 3, padding='same', kernel_initializer='he_normal', 100 kernel_regularizer=l2(1e-4))(residual) 101 102 # Downsampling 103 if downsample_strides > 1: 104 identity = AveragePooling2D(pool_size=(1,1), strides=(2,2))(identity) 105 106 # Zero_padding to match channels 107 if in_channels != out_channels: 108 zeros_identity = keras.layers.subtract([identity, identity]) 109 identity = keras.layers.concatenate([identity, zeros_identity]) 110 in_channels = out_channels 111 112 residual = keras.layers.add([residual, identity]) 113 114 return residual 115 116 117 # define and train a model 118 inputs = Input(shape=(32, 32, 3)) 119 net = Conv2D(16, 3, padding='same', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(inputs) 120 net = residual_block(net, 3, 16, downsample=False) 121 net = residual_block(net, 1, 32, downsample=True) 122 net = residual_block(net, 2, 32, downsample=False) 123 net = residual_block(net, 1, 64, downsample=True) 124 net = residual_block(net, 2, 64, downsample=False) 125 net = BatchNormalization(momentum=0.9, gamma_regularizer=l2(1e-4))(net) 126 net = Activation('relu')(net) 127 net = GlobalAveragePooling2D()(net) 128 outputs = Dense(10, activation='softmax', kernel_initializer='he_normal', kernel_regularizer=l2(1e-4))(net) 129 model = Model(inputs=inputs, outputs=outputs) 130 sgd = optimizers.SGD(lr=0.1, decay=0., momentum=0.9, nesterov=True) 131 model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy']) 132 133 # data augmentation 134 datagen = ImageDataGenerator( 135 # randomly rotate images in the range (deg 0 to 180) 136 rotation_range=30, 137 # Range for random zoom 138 zoom_range = 0.2, 139 # shear angle in counter-clockwise direction in degrees 140 shear_range = 30, 141 # randomly flip images 142 horizontal_flip=True, 143 # randomly shift images horizontally 144 width_shift_range=0.125, 145 # randomly shift images vertically 146 height_shift_range=0.125) 147 148 reduce_lr = LearningRateScheduler(scheduler) 149 # fit the model on the batches generated by datagen.flow(). 150 model.fit_generator(datagen.flow(x_train, y_train, batch_size=625), 151 validation_data=(x_test, y_test), epochs=5000, 152 verbose=1, callbacks=[reduce_lr], workers=10) 153 154 # get results 155 K.set_learning_phase(0) 156 DRSN_train_score = model.evaluate(x_train, y_train, batch_size=625, verbose=0) 157 print('Train loss:', DRSN_train_score[0]) 158 print('Train accuracy:', DRSN_train_score[1]) 159 DRSN_test_score = model.evaluate(x_test, y_test, batch_size=625, verbose=0) 160 print('Test loss:', DRSN_test_score[0]) 161 print('Test accuracy:', DRSN_test_score[1])
实验结果如下：
1 Epoch 2500/5000 2 12s 151ms/step - loss: 0.1258 - acc: 0.9867 - val_loss: 0.4697 - val_acc: 0.9024 3 Epoch 2501/5000 4 12s 151ms/step - loss: 0.1274 - acc: 0.9852 - val_loss: 0.4688 - val_acc: 0.9026 5 Epoch 2502/5000 6 12s 151ms/step - loss: 0.1260 - acc: 0.9861 - val_loss: 0.4585 - val_acc: 0.9040 7 Epoch 2503/5000 8 12s 152ms/step - loss: 0.1241 - acc: 0.9869 - val_loss: 0.4489 - val_acc: 0.9066 9 Epoch 2504/5000 10 12s 152ms/step - loss: 0.1236 - acc: 0.9869 - val_loss: 0.4469 - val_acc: 0.9106 11 Epoch 2505/5000 12 12s 151ms/step - loss: 0.1276 - acc: 0.9850 - val_loss: 0.4515 - val_acc: 0.9034 13 Epoch 2506/5000 14 12s 151ms/step - loss: 0.1252 - acc: 0.9864 - val_loss: 0.4586 - val_acc: 0.9074 15 Epoch 2507/5000 16 12s 151ms/step - loss: 0.1289 - acc: 0.9852 - val_loss: 0.4585 - val_acc: 0.9057 17 Epoch 2508/5000 18 12s 151ms/step - loss: 0.1285 - acc: 0.9853 - val_loss: 0.4485 - val_acc: 0.9077 19 Epoch 2509/5000 20 12s 151ms/step - loss: 0.1284 - acc: 0.9851 - val_loss: 0.4529 - val_acc: 0.9032 21 Epoch 2510/5000 22 12s 151ms/step - loss: 0.1287 - acc: 0.9855 - val_loss: 0.4567 - val_acc: 0.9040 23 Epoch 2511/5000 24 12s 151ms/step - loss: 0.1253 - acc: 0.9862 - val_loss: 0.4554 - val_acc: 0.9080 25 Epoch 2512/5000 26 12s 152ms/step - loss: 0.1262 - acc: 0.9859 - val_loss: 0.4477 - val_acc: 0.9086 27 Epoch 2513/5000 28 12s 151ms/step - loss: 0.1241 - acc: 0.9864 - val_loss: 0.4531 - val_acc: 0.9063 29 Epoch 2514/5000 30 12s 150ms/step - loss: 0.1247 - acc: 0.9866 - val_loss: 0.4484 - val_acc: 0.9073 31 Epoch 2515/5000 32 12s 151ms/step - loss: 0.1239 - acc: 0.9869 - val_loss: 0.4502 - val_acc: 0.9078 33 Epoch 2516/5000 34 12s 151ms/step - loss: 0.1275 - acc: 0.9857 - val_loss: 0.4790 - val_acc: 0.8981 35 Epoch 2517/5000 36 12s 152ms/step - loss: 0.1259 - acc: 0.9862 - val_loss: 0.4625 - val_acc: 0.9063 37 Epoch 2518/5000 38 12s 151ms/step - loss: 0.1278 - acc: 0.9853 - val_loss: 0.4751 - val_acc: 0.9009 39 Epoch 2519/5000 40 12s 151ms/step - loss: 0.1283 - acc: 0.9857 - val_loss: 0.4655 - val_acc: 0.9056 41 Epoch 2520/5000 42 12s 151ms/step - loss: 0.1275 - acc: 0.9859 - val_loss: 0.4386 - val_acc: 0.9085 43 Epoch 2521/5000 44 12s 151ms/step - loss: 0.1245 - acc: 0.9871 - val_loss: 0.4699 - val_acc: 0.9006 45 Epoch 2522/5000 46 12s 151ms/step - loss: 0.1278 - acc: 0.9860 - val_loss: 0.4520 - val_acc: 0.9050 47 Epoch 2523/5000 48 12s 151ms/step - loss: 0.1249 - acc: 0.9864 - val_loss: 0.4566 - val_acc: 0.9056 49 Epoch 2524/5000 50 12s 152ms/step - loss: 0.1278 - acc: 0.9855 - val_loss: 0.4650 - val_acc: 0.9018 51 Epoch 2525/5000 52 12s 151ms/step - loss: 0.1235 - acc: 0.9873 - val_loss: 0.4555 - val_acc: 0.9061 53 Epoch 2526/5000 54 12s 151ms/step - loss: 0.1260 - acc: 0.9862 - val_loss: 0.4556 - val_acc: 0.9061 55 Epoch 2527/5000 56 12s 152ms/step - loss: 0.1261 - acc: 0.9866 - val_loss: 0.4667 - val_acc: 0.9040 57 Epoch 2528/5000 58 12s 152ms/step - loss: 0.1240 - acc: 0.9874 - val_loss: 0.4539 - val_acc: 0.9083 59 Epoch 2529/5000 60 12s 152ms/step - loss: 0.1281 - acc: 0.9856 - val_loss: 0.4584 - val_acc: 0.9048 61 Epoch 2530/5000 62 12s 151ms/step - loss: 0.1234 - acc: 0.9871 - val_loss: 0.4538 - val_acc: 0.9048 63 Epoch 2531/5000 64 12s 151ms/step - loss: 0.1235 - acc: 0.9868 - val_loss: 0.4504 - val_acc: 0.9056 65 Epoch 2532/5000 66 12s 151ms/step - loss: 0.1247 - acc: 0.9871 - val_loss: 0.4529 - val_acc: 0.9053 67 Epoch 2533/5000 68 12s 150ms/step - loss: 0.1241 - acc: 0.9872 - val_loss: 0.4591 - val_acc: 0.9034 69 Epoch 2534/5000 70 12s 152ms/step - loss: 0.1255 - acc: 0.9865 - val_loss: 0.4502 - val_acc: 0.9058 71 Epoch 2535/5000 72 12s 151ms/step - loss: 0.1254 - acc: 0.9865 - val_loss: 0.4596 - val_acc: 0.9039 73 Epoch 2536/5000 74 12s 152ms/step - loss: 0.1239 - acc: 0.9872 - val_loss: 0.4488 - val_acc: 0.9040 75 Epoch 2537/5000 76 12s 151ms/step - loss: 0.1260 - acc: 0.9865 - val_loss: 0.4494 - val_acc: 0.9042 77 Epoch 2538/5000 78 12s 150ms/step - loss: 0.1288 - acc: 0.9851 - val_loss: 0.4621 - val_acc: 0.9039 79 Epoch 2539/5000 80 12s 152ms/step - loss: 0.1267 - acc: 0.9855 - val_loss: 0.4497 - val_acc: 0.9068 81 Epoch 2540/5000 82 12s 151ms/step - loss: 0.1250 - acc: 0.9869 - val_loss: 0.4626 - val_acc: 0.9024 83 Epoch 2541/5000 84 12s 152ms/step - loss: 0.1272 - acc: 0.9856 - val_loss: 0.4621 - val_acc: 0.9038 85 Epoch 2542/5000 86 12s 151ms/step - loss: 0.1258 - acc: 0.9862 - val_loss: 0.4738 - val_acc: 0.9044 87 Epoch 2543/5000 88 12s 152ms/step - loss: 0.1257 - acc: 0.9862 - val_loss: 0.4597 - val_acc: 0.9061 89 Epoch 2544/5000 90 12s 151ms/step - loss: 0.1271 - acc: 0.9854 - val_loss: 0.4571 - val_acc: 0.9008 91 Epoch 2545/5000 92 12s 151ms/step - loss: 0.1247 - acc: 0.9861 - val_loss: 0.4450 - val_acc: 0.9065 93 Epoch 2546/5000 94 12s 152ms/step - loss: 0.1273 - acc: 0.9860 - val_loss: 0.4568 - val_acc: 0.9031 95 Epoch 2547/5000 96 12s 151ms/step - loss: 0.1291 - acc: 0.9855 - val_loss: 0.4558 - val_acc: 0.9034 97 Epoch 2548/5000 98 12s 152ms/step - loss: 0.1280 - acc: 0.9849 - val_loss: 0.4463 - val_acc: 0.9077 99 Epoch 2549/5000 100 12s 151ms/step - loss: 0.1237 - acc: 0.9868 - val_loss: 0.4427 - val_acc: 0.9083 101 Epoch 2550/5000 102 12s 151ms/step - loss: 0.1247 - acc: 0.9865 - val_loss: 0.4486 - val_acc: 0.9060 103 Epoch 2551/5000 104 12s 152ms/step - loss: 0.1265 - acc: 0.9864 - val_loss: 0.4414 - val_acc: 0.9047 105 Epoch 2552/5000 106 12s 151ms/step - loss: 0.1275 - acc: 0.9859 - val_loss: 0.4652 - val_acc: 0.9003 107 Epoch 2553/5000 108 12s 151ms/step - loss: 0.1241 - acc: 0.9864 - val_loss: 0.4713 - val_acc: 0.8976 109 Epoch 2554/5000 110 12s 152ms/step - loss: 0.1258 - acc: 0.9862 - val_loss: 0.4549 - val_acc: 0.9048 111 Epoch 2555/5000 112 12s 151ms/step - loss: 0.1249 - acc: 0.9866 - val_loss: 0.4376 - val_acc: 0.9069 113 Epoch 2556/5000 114 12s 152ms/step - loss: 0.1251 - acc: 0.9866 - val_loss: 0.4519 - val_acc: 0.9062 115 Epoch 2557/5000 116 12s 151ms/step - loss: 0.1269 - acc: 0.9857 - val_loss: 0.4479 - val_acc: 0.9069 117 Epoch 2558/5000 118 12s 151ms/step - loss: 0.1240 - acc: 0.9870 - val_loss: 0.4629 - val_acc: 0.9023 119 Epoch 2559/5000 120 12s 151ms/step - loss: 0.1257 - acc: 0.9866 - val_loss: 0.4487 - val_acc: 0.9039 121 Epoch 2560/5000 122 12s 151ms/step - loss: 0.1272 - acc: 0.9859 - val_loss: 0.4574 - val_acc: 0.9029 123 Epoch 2561/5000 124 12s 152ms/step - loss: 0.1238 - acc: 0.9872 - val_loss: 0.4530 - val_acc: 0.9073 125 Epoch 2562/5000 126 12s 152ms/step - loss: 0.1226 - acc: 0.9872 - val_loss: 0.4589 - val_acc: 0.9048 127 Epoch 2563/5000 128 12s 151ms/step - loss: 0.1283 - acc: 0.9854 - val_loss: 0.4525 - val_acc: 0.9032 129 Epoch 2564/5000 130 12s 151ms/step - loss: 0.1286 - acc: 0.9851 - val_loss: 0.4488 - val_acc: 0.9063 131 Epoch 2565/5000 132 12s 150ms/step - loss: 0.1263 - acc: 0.9862 - val_loss: 0.4520 - val_acc: 0.9044 133 Epoch 2566/5000 134 12s 152ms/step - loss: 0.1280 - acc: 0.9854 - val_loss: 0.4561 - val_acc: 0.9025 135 Epoch 2567/5000 136 12s 151ms/step - loss: 0.1259 - acc: 0.9860 - val_loss: 0.4532 - val_acc: 0.9034 137 Epoch 2568/5000 138 12s 156ms/step - loss: 0.1249 - acc: 0.9864 - val_loss: 0.4449 - val_acc: 0.9072 139 Epoch 2569/5000 140 12s 152ms/step - loss: 0.1269 - acc: 0.9857 - val_loss: 0.4465 - val_acc: 0.9056 141 Epoch 2570/5000 142 12s 153ms/step - loss: 0.1282 - acc: 0.9853 - val_loss: 0.4445 - val_acc: 0.9074 143 Epoch 2571/5000 144 12s 153ms/step - loss: 0.1268 - acc: 0.9857 - val_loss: 0.4496 - val_acc: 0.9028 145 Epoch 2572/5000 146 12s 152ms/step - loss: 0.1255 - acc: 0.9860 - val_loss: 0.4600 - val_acc: 0.9038 147 Epoch 2573/5000 148 12s 153ms/step - loss: 0.1206 - acc: 0.9884 - val_loss: 0.4555 - val_acc: 0.9057 149 Epoch 2574/5000 150 12s 152ms/step - loss: 0.1242 - acc: 0.9867 - val_loss: 0.4483 - val_acc: 0.9071 151 Epoch 2575/5000 152 12s 153ms/step - loss: 0.1225 - acc: 0.9871 - val_loss: 0.4497 - val_acc: 0.9054 153 Epoch 2576/5000 154 12s 152ms/step - loss: 0.1233 - acc: 0.9876 - val_loss: 0.4645 - val_acc: 0.9039 155 Epoch 2577/5000 156 12s 153ms/step - loss: 0.1247 - acc: 0.9865 - val_loss: 0.4584 - val_acc: 0.9036 157 Epoch 2578/5000 158 12s 153ms/step - loss: 0.1248 - acc: 0.9864 - val_loss: 0.4666 - val_acc: 0.9045 159 Epoch 2579/5000 160 12s 153ms/step - loss: 0.1245 - acc: 0.9868 - val_loss: 0.4668 - val_acc: 0.9063 161 Epoch 2580/5000 162 12s 153ms/step - loss: 0.1253 - acc: 0.9862 - val_loss: 0.4609 - val_acc: 0.9023 163 Epoch 2581/5000 164 12s 153ms/step - loss: 0.1242 - acc: 0.9869 - val_loss: 0.4450 - val_acc: 0.9058 165 Epoch 2582/5000 166 12s 152ms/step - loss: 0.1253 - acc: 0.9863 - val_loss: 0.4391 - val_acc: 0.9068 167 Epoch 2583/5000 168 12s 153ms/step - loss: 0.1266 - acc: 0.9861 - val_loss: 0.4420 - val_acc: 0.9066 169 Epoch 2584/5000 170 12s 153ms/step - loss: 0.1255 - acc: 0.9865 - val_loss: 0.4480 - val_acc: 0.9056 171 Epoch 2585/5000 172 12s 152ms/step - loss: 0.1281 - acc: 0.9851 - val_loss: 0.4449 - val_acc: 0.9052 173 Epoch 2586/5000 174 12s 152ms/step - loss: 0.1247 - acc: 0.9868 - val_loss: 0.4536 - val_acc: 0.9050 175 Epoch 2587/5000 176 12s 152ms/step - loss: 0.1273 - acc: 0.9857 - val_loss: 0.4712 - val_acc: 0.9007 177 Epoch 2588/5000 178 12s 153ms/step - loss: 0.1292 - acc: 0.9852 - val_loss: 0.4495 - val_acc: 0.9059 179 Epoch 2589/5000 180 12s 153ms/step - loss: 0.1253 - acc: 0.9866 - val_loss: 0.4626 - val_acc: 0.9051 181 Epoch 2590/5000 182 12s 153ms/step - loss: 0.1248 - acc: 0.9867 - val_loss: 0.4609 - val_acc: 0.9021 183 Epoch 2591/5000 184 12s 152ms/step - loss: 0.1273 - acc: 0.9855 - val_loss: 0.4594 - val_acc: 0.9039 185 Epoch 2592/5000 186 12s 152ms/step - loss: 0.1257 - acc: 0.9857 - val_loss: 0.4519 - val_acc: 0.9023 187 Epoch 2593/5000 188 12s 152ms/step - loss: 0.1317 - acc: 0.9845 - val_loss: 0.4526 - val_acc: 0.9063 189 Epoch 2594/5000 190 12s 153ms/step - loss: 0.1255 - acc: 0.9864 - val_loss: 0.4529 - val_acc: 0.9066 191 Epoch 2595/5000 192 12s 153ms/step - loss: 0.1244 - acc: 0.9863 - val_loss: 0.4540 - val_acc: 0.9076 193 Epoch 2596/5000 194 12s 153ms/step - loss: 0.1268 - acc: 0.9859 - val_loss: 0.4632 - val_acc: 0.9022 195 Epoch 2597/5000 196 12s 153ms/step - loss: 0.1250 - acc: 0.9864 - val_loss: 0.4440 - val_acc: 0.9057 197 Epoch 2598/5000 198 12s 153ms/step - loss: 0.1246 - acc: 0.9870 - val_loss: 0.4489 - val_acc: 0.9035 199 Epoch 2599/5000 200 12s 153ms/step - loss: 0.1252 - acc: 0.9857 - val_loss: 0.4671 - val_acc: 0.9035 201 Epoch 2600/5000 202 12s 153ms/step - loss: 0.1253 - acc: 0.9866 - val_loss: 0.4532 - val_acc: 0.9077 203 Epoch 2601/5000 204 12s 153ms/step - loss: 0.1228 - acc: 0.9870 - val_loss: 0.4503 - val_acc: 0.9026 205 Epoch 2602/5000 206 12s 153ms/step - loss: 0.1225 - acc: 0.9873 - val_loss: 0.4490 - val_acc: 0.9027 207 Epoch 2603/5000 208 12s 152ms/step - loss: 0.1238 - acc: 0.9871 - val_loss: 0.4430 - val_acc: 0.9066 209 Epoch 2604/5000 210 12s 152ms/step - loss: 0.1279 - acc: 0.9856 - val_loss: 0.4576 - val_acc: 0.9054 211 Epoch 2605/5000 212 12s 152ms/step - loss: 0.1253 - acc: 0.9864 - val_loss: 0.4425 - val_acc: 0.9069 213 Epoch 2606/5000 214 12s 152ms/step - loss: 0.1269 - acc: 0.9859 - val_loss: 0.4542 - val_acc: 0.9024 215 Epoch 2607/5000 216 12s 152ms/step - loss: 0.1281 - acc: 0.9852 - val_loss: 0.4673 - val_acc: 0.9023 217 Epoch 2608/5000 218 12s 152ms/step - loss: 0.1269 - acc: 0.9864 - val_loss: 0.4638 - val_acc: 0.9025 219 Epoch 2609/5000 220 12s 152ms/step - loss: 0.1261 - acc: 0.9861 - val_loss: 0.4499 - val_acc: 0.9059 221 Epoch 2610/5000 222 12s 152ms/step - loss: 0.1240 - acc: 0.9871 - val_loss: 0.4502 - val_acc: 0.9070 223 Epoch 2611/5000 224 12s 151ms/step - loss: 0.1236 - acc: 0.9874 - val_loss: 0.4592 - val_acc: 0.9018 225 Epoch 2612/5000 226 12s 151ms/step - loss: 0.1233 - acc: 0.9874 - val_loss: 0.4603 - val_acc: 0.9032 227 Epoch 2613/5000 228 12s 151ms/step - loss: 0.1265 - acc: 0.9853 - val_loss: 0.4574 - val_acc: 0.9056 229 Epoch 2614/5000 230 12s 152ms/step - loss: 0.1229 - acc: 0.9871 - val_loss: 0.4514 - val_acc: 0.9052 231 Epoch 2615/5000 232 12s 152ms/step - loss: 0.1233 - acc: 0.9869 - val_loss: 0.4699 - val_acc: 0.9013 233 Epoch 2616/5000 234 12s 151ms/step - loss: 0.1248 - acc: 0.9863 - val_loss: 0.4715 - val_acc: 0.8995 235 Epoch 2617/5000 236 12s 151ms/step - loss: 0.1284 - acc: 0.9853 - val_loss: 0.4647 - val_acc: 0.9043 237 Epoch 2618/5000 238 12s 151ms/step - loss: 0.1267 - acc: 0.9857 - val_loss: 0.4656 - val_acc: 0.9005 239 Epoch 2619/5000 240 12s 152ms/step - loss: 0.1232 - acc: 0.9874 - val_loss: 0.4657 - val_acc: 0.9035 241 Epoch 2620/5000 242 12s 152ms/step - loss: 0.1274 - acc: 0.9859 - val_loss: 0.4522 - val_acc: 0.9051 243 Epoch 2621/5000 244 12s 151ms/step - loss: 0.1275 - acc: 0.9859 - val_loss: 0.4528 - val_acc: 0.9034 245 Epoch 2622/5000 246 12s 152ms/step - loss: 0.1258 - acc: 0.9863 - val_loss: 0.4600 - val_acc: 0.9036 247 Epoch 2623/5000 248 12s 152ms/step - loss: 0.1245 - acc: 0.9865 - val_loss: 0.4626 - val_acc: 0.9047 249 Epoch 2624/5000 250 12s 152ms/step - loss: 0.1241 - acc: 0.9866 - val_loss: 0.4644 - val_acc: 0.9043 251 Epoch 2625/5000 252 12s 152ms/step - loss: 0.1245 - acc: 0.9871 - val_loss: 0.4762 - val_acc: 0.9035 253 Epoch 2626/5000 254 12s 152ms/step - loss: 0.1263 - acc: 0.9859 - val_loss: 0.4579 - val_acc: 0.9033 255 Epoch 2627/5000 256 12s 151ms/step - loss: 0.1253 - acc: 0.9867 - val_loss: 0.4616 - val_acc: 0.9022 257 Epoch 2628/5000 258 12s 151ms/step - loss: 0.1268 - acc: 0.9858 - val_loss: 0.4721 - val_acc: 0.9026 259 Epoch 2629/5000 260 12s 151ms/step - loss: 0.1270 - acc: 0.9854 - val_loss: 0.4528 - val_acc: 0.9048 261 Epoch 2630/5000 262 12s 151ms/step - loss: 0.1258 - acc: 0.9863 - val_loss: 0.4496 - val_acc: 0.9056 263 Epoch 2631/5000 264 12s 152ms/step - loss: 0.1241 - acc: 0.9868 - val_loss: 0.4469 - val_acc: 0.9058 265 Epoch 2632/5000 266 12s 151ms/step - loss: 0.1261 - acc: 0.9865 - val_loss: 0.4923 - val_acc: 0.8972 267 Epoch 2633/5000 268 12s 152ms/step - loss: 0.1255 - acc: 0.9860 - val_loss: 0.4662 - val_acc: 0.9011 269 Epoch 2634/5000 270 12s 151ms/step - loss: 0.1230 - acc: 0.9873 - val_loss: 0.4461 - val_acc: 0.9055 271 Epoch 2635/5000 272 12s 151ms/step - loss: 0.1206 - acc: 0.9877 - val_loss: 0.4495 - val_acc: 0.9055 273 Epoch 2636/5000 274 12s 152ms/step - loss: 0.1234 - acc: 0.9874 - val_loss: 0.4671 - val_acc: 0.9053 275 Epoch 2637/5000 276 12s 152ms/step - loss: 0.1233 - acc: 0.9872 - val_loss: 0.4637 - val_acc: 0.9032 277 Epoch 2638/5000 278 12s 151ms/step - loss: 0.1221 - acc: 0.9874 - val_loss: 0.4634 - val_acc: 0.9042 279 Epoch 2639/5000 280 12s 151ms/step - loss: 0.1209 - acc: 0.9877 - val_loss: 0.4655 - val_acc: 0.9023 281 Epoch 2640/5000 282 12s 152ms/step - loss: 0.1258 - acc: 0.9864 - val_loss: 0.4556 - val_acc: 0.9065 283 Epoch 2641/5000 284 12s 152ms/step - loss: 0.1247 - acc: 0.9867 - val_loss: 0.4576 - val_acc: 0.9018 285 Epoch 2642/5000 286 12s 152ms/step - loss: 0.1274 - acc: 0.9855 - val_loss: 0.4584 - val_acc: 0.9051 287 Epoch 2643/5000 288 12s 152ms/step - loss: 0.1282 - acc: 0.9856 - val_loss: 0.4528 - val_acc: 0.9066 289 Epoch 2644/5000 290 12s 151ms/step - loss: 0.1270 - acc: 0.9858 - val_loss: 0.4617 - val_acc: 0.9015 291 Epoch 2645/5000 292 12s 152ms/step - loss: 0.1279 - acc: 0.9853 - val_loss: 0.4448 - val_acc: 0.9063 293 Epoch 2646/5000 294 12s 151ms/step - loss: 0.1256 - acc: 0.9865 - val_loss: 0.4449 - val_acc: 0.9055 295 Epoch 2647/5000 296 12s 152ms/step - loss: 0.1259 - acc: 0.9864 - val_loss: 0.4429 - val_acc: 0.9052 297 Epoch 2648/5000 298 12s 152ms/step - loss: 0.1244 - acc: 0.9869 - val_loss: 0.4474 - val_acc: 0.9038 299 Epoch 2649/5000 300 12s 151ms/step - loss: 0.1236 - acc: 0.9874 - val_loss: 0.4459 - val_acc: 0.9072 301 Epoch 2650/5000 302 12s 151ms/step - loss: 0.1246 - acc: 0.9872 - val_loss: 0.4469 - val_acc: 0.9039 303 Epoch 2651/5000 304 12s 151ms/step - loss: 0.1254 - acc: 0.9868 - val_loss: 0.4540 - val_acc: 0.9056 305 Epoch 2652/5000 306 12s 151ms/step - loss: 0.1261 - acc: 0.9866 - val_loss: 0.4616 - val_acc: 0.9003 307 Epoch 2653/5000 308 12s 151ms/step - loss: 0.1254 - acc: 0.9860 - val_loss: 0.4525 - val_acc: 0.9029 309 Epoch 2654/5000 310 12s 151ms/step - loss: 0.1226 - acc: 0.9874 - val_loss: 0.4589 - val_acc: 0.9032 311 Epoch 2655/5000 312 12s 151ms/step - loss: 0.1244 - acc: 0.9868 - val_loss: 0.4548 - val_acc: 0.9027 313 Epoch 2656/5000 314 12s 151ms/step - loss: 0.1252 - acc: 0.9871 - val_loss: 0.4438 - val_acc: 0.9057 315 Epoch 2657/5000 316 12s 151ms/step - loss: 0.1228 - acc: 0.9869 - val_loss: 0.4554 - val_acc: 0.9045 317 Epoch 2658/5000 318 12s 152ms/step - loss: 0.1280 - acc: 0.9857 - val_loss: 0.4481 - val_acc: 0.9066 319 Epoch 2659/5000 320 12s 152ms/step - loss: 0.1251 - acc: 0.9861 - val_loss: 0.4492 - val_acc: 0.9075 321 Epoch 2660/5000 322 12s 151ms/step - loss: 0.1222 - acc: 0.9873 - val_loss: 0.4501 - val_acc: 0.9045 323 Epoch 2661/5000 324 12s 152ms/step - loss: 0.1251 - acc: 0.9864 - val_loss: 0.4597 - val_acc: 0.9040 325 Epoch 2662/5000 326 12s 151ms/step - loss: 0.1258 - acc: 0.9860 - val_loss: 0.4588 - val_acc: 0.9039 327 Epoch 2663/5000 328 12s 152ms/step - loss: 0.1235 - acc: 0.9863 - val_loss: 0.4472 - val_acc: 0.9056 329 Epoch 2664/5000 330 12s 152ms/step - loss: 0.1215 - acc: 0.9874 - val_loss: 0.4674 - val_acc: 0.9004 331 Epoch 2665/5000 332 12s 151ms/step - loss: 0.1239 - acc: 0.9864 - val_loss: 0.4674 - val_acc: 0.9026 333 Epoch 2666/5000 334 12s 151ms/step - loss: 0.1241 - acc: 0.9867 - val_loss: 0.4636 - val_acc: 0.9023 335 Epoch 2667/5000 336 12s 151ms/step - loss: 0.1250 - acc: 0.9866 - val_loss: 0.4620 - val_acc: 0.9025 337 Epoch 2668/5000 338 12s 151ms/step - loss: 0.1252 - acc: 0.9864 - val_loss: 0.4758 - val_acc: 0.8995 339 Epoch 2669/5000 340 12s 152ms/step - loss: 0.1278 - acc: 0.9858 - val_loss: 0.4816 - val_acc: 0.8986 341 Epoch 2670/5000 342 12s 152ms/step - loss: 0.1251 - acc: 0.9864 - val_loss: 0.4692 - val_acc: 0.9009 343 Epoch 2671/5000 344 12s 151ms/step - loss: 0.1281 - acc: 0.9852 - val_loss: 0.4615 - val_acc: 0.9024 345 Epoch 2672/5000 346 12s 152ms/step - loss: 0.1233 - acc: 0.9868 - val_loss: 0.4583 - val_acc: 0.9055 347 Epoch 2673/5000 348 12s 152ms/step - loss: 0.1228 - acc: 0.9871 - val_loss: 0.4689 - val_acc: 0.9039 349 Epoch 2674/5000 350 12s 152ms/step - loss: 0.1267 - acc: 0.9856 - val_loss: 0.4596 - val_acc: 0.9049 351 Epoch 2675/5000 352 12s 151ms/step - loss: 0.1289 - acc: 0.9847 - val_loss: 0.4575 - val_acc: 0.9020 353 Epoch 2676/5000 354 12s 152ms/step - loss: 0.1234 - acc: 0.9870 - val_loss: 0.4527 - val_acc: 0.9068 355 Epoch 2677/5000 356 12s 151ms/step - loss: 0.1248 - acc: 0.9865 - val_loss: 0.4588 - val_acc: 0.9035 357 Epoch 2678/5000 358 12s 151ms/step - loss: 0.1234 - acc: 0.9866 - val_loss: 0.4667 - val_acc: 0.9009 359 Epoch 2679/5000 360 12s 152ms/step - loss: 0.1234 - acc: 0.9869 - val_loss: 0.4613 - val_acc: 0.9032 361 Epoch 2680/5000 362 12s 151ms/step - loss: 0.1248 - acc: 0.9860 - val_loss: 0.4748 - val_acc: 0.9014 363 Epoch 2681/5000 364 12s 152ms/step - loss: 0.1256 - acc: 0.9856 - val_loss: 0.4579 - val_acc: 0.9051 365 Epoch 2682/5000 366 12s 151ms/step - loss: 0.1276 - acc: 0.9854 - val_loss: 0.4688 - val_acc: 0.9019 367 Epoch 2683/5000 368 12s 152ms/step - loss: 0.1237 - acc: 0.9866 - val_loss: 0.4623 - val_acc: 0.9023 369 Epoch 2684/5000 370 12s 152ms/step - loss: 0.1232 - acc: 0.9872 - val_loss: 0.4618 - val_acc: 0.9033 371 Epoch 2685/5000 372 12s 151ms/step - loss: 0.1253 - acc: 0.9865 - val_loss: 0.4712 - val_acc: 0.9007 373 Epoch 2686/5000 374 12s 151ms/step - loss: 0.1248 - acc: 0.9864 - val_loss: 0.4675 - val_acc: 0.9035 375 Epoch 2687/5000 376 12s 152ms/step - loss: 0.1291 - acc: 0.9851 - val_loss: 0.4600 - val_acc: 0.9031 377 Epoch 2688/5000 378 12s 151ms/step - loss: 0.1255 - acc: 0.9862 - val_loss: 0.4623 - val_acc: 0.9017 379 Epoch 2689/5000 380 12s 152ms/step - loss: 0.1273 - acc: 0.9854 - val_loss: 0.4609 - val_acc: 0.9021 381 Epoch 2690/5000 382 12s 152ms/step - loss: 0.1262 - acc: 0.9862 - val_loss: 0.4454 - val_acc: 0.9048 383 Epoch 2691/5000 384 12s 151ms/step - loss: 0.1231 - acc: 0.9869 - val_loss: 0.4612 - val_acc: 0.9040 385 Epoch 2692/5000 386 12s 151ms/step - loss: 0.1254 - acc: 0.9867 - val_loss: 0.4524 - val_acc: 0.9045 387 Epoch 2693/5000 388 12s 152ms/step - loss: 0.1233 - acc: 0.9874 - val_loss: 0.4567 - val_acc: 0.9045 389 Epoch 2694/5000 390 12s 151ms/step - loss: 0.1243 - acc: 0.9864 - val_loss: 0.4603 - val_acc: 0.9023 391 Epoch 2695/5000 392 12s 151ms/step - loss: 0.1269 - acc: 0.9861 - val_loss: 0.4714 - val_acc: 0.8998 393 Epoch 2696/5000 394 12s 152ms/step - loss: 0.1240 - acc: 0.9866 - val_loss: 0.4402 - val_acc: 0.9068 395 Epoch 2697/5000 396 12s 156ms/step - loss: 0.1245 - acc: 0.9864 - val_loss: 0.4597 - val_acc: 0.9040 397 Epoch 2698/5000 398 12s 151ms/step - loss: 0.1255 - acc: 0.9863 - val_loss: 0.4499 - val_acc: 0.9045 399 Epoch 2699/5000 400 12s 152ms/step - loss: 0.1223 - acc: 0.9876 - val_loss: 0.4660 - val_acc: 0.9054 401 Epoch 2700/5000 402 12s 152ms/step - loss: 0.1248 - acc: 0.9865 - val_loss: 0.4537 - val_acc: 0.9045 403 Epoch 2701/5000 404 12s 154ms/step - loss: 0.1252 - acc: 0.9864 - val_loss: 0.4683 - val_acc: 0.9019 405 Epoch 2702/5000 406 12s 153ms/step - loss: 0.1254 - acc: 0.9859 - val_loss: 0.4657 - val_acc: 0.9039 407 Epoch 2703/5000 408 12s 153ms/step - loss: 0.1234 - acc: 0.9874 - val_loss: 0.4679 - val_acc: 0.9006 409 Epoch 2704/5000 410 12s 153ms/step - loss: 0.1268 - acc: 0.9856 - val_loss: 0.4724 - val_acc: 0.8994 411 Epoch 2705/5000 412 12s 153ms/step - loss: 0.1244 - acc: 0.9869 - val_loss: 0.4762 - val_acc: 0.8988 413 Epoch 2706/5000 414 12s 153ms/step - loss: 0.1245 - acc: 0.9869 - val_loss: 0.4669 - val_acc: 0.9034 415 Epoch 2707/5000 416 12s 153ms/step - loss: 0.1226 - acc: 0.9873 - val_loss: 0.4629 - val_acc: 0.9046 417 Epoch 2708/5000 418 12s 152ms/step - loss: 0.1244 - acc: 0.9868 - val_loss: 0.4528 - val_acc: 0.9066 419 Epoch 2709/5000 420 12s 153ms/step - loss: 0.1208 - acc: 0.9874 - val_loss: 0.4600 - val_acc: 0.9013 421 Epoch 2710/5000 422 12s 153ms/step - loss: 0.1251 - acc: 0.9856 - val_loss: 0.4551 - val_acc: 0.9039 423 Epoch 2711/5000 424 12s 153ms/step - loss: 0.1242 - acc: 0.9872 - val_loss: 0.4457 - val_acc: 0.9073 425 Epoch 2712/5000 426 12s 153ms/step - loss: 0.1269 - acc: 0.9859 - val_loss: 0.4577 - val_acc: 0.9027 427 Epoch 2713/5000 428 12s 153ms/step - loss: 0.1295 - acc: 0.9846 - val_loss: 0.4609 - val_acc: 0.9039 429 Epoch 2714/5000 430 12s 153ms/step - loss: 0.1244 - acc: 0.9870 - val_loss: 0.4614 - val_acc: 0.9019 431 Epoch 2715/5000 432 12s 153ms/step - loss: 0.1213 - acc: 0.9877 - val_loss: 0.4560 - val_acc: 0.9046 433 Epoch 2716/5000 434 12s 152ms/step - loss: 0.1252 - acc: 0.9862 - val_loss: 0.4501 - val_acc: 0.9059 435 Epoch 2717/5000 436 12s 153ms/step - loss: 0.1257 - acc: 0.9860 - val_loss: 0.4686 - val_acc: 0.9015 437 Epoch 2718/5000 438 12s 153ms/step - loss: 0.1233 - acc: 0.9870 - val_loss: 0.4636 - val_acc: 0.9022 439 Epoch 2719/5000 440 12s 153ms/step - loss: 0.1242 - acc: 0.9864 - val_loss: 0.4403 - val_acc: 0.9086 441 Epoch 2720/5000 442 12s 153ms/step - loss: 0.1268 - acc: 0.9858 - val_loss: 0.4516 - val_acc: 0.9050 443 Epoch 2721/5000 444 12s 152ms/step - loss: 0.1222 - acc: 0.9876 - val_loss: 0.4555 - val_acc: 0.9055 445 Epoch 2722/5000 446 12s 152ms/step - loss: 0.1192 - acc: 0.9883 - val_loss: 0.4387 - val_acc: 0.9076 447 Epoch 2723/5000 448 12s 152ms/step - loss: 0.1235 - acc: 0.9868 - val_loss: 0.4663 - val_acc: 0.9059 449 Epoch 2724/5000 450 12s 152ms/step - loss: 0.1246 - acc: 0.9862 - val_loss: 0.4729 - val_acc: 0.9028 451 Epoch 2725/5000 452 12s 152ms/step - loss: 0.1291 - acc: 0.9844 - val_loss: 0.4582 - val_acc: 0.9037 453 Epoch 2726/5000 454 12s 152ms/step - loss: 0.1228 - acc: 0.9872 - val_loss: 0.4613 - val_acc: 0.9028 455 Epoch 2727/5000 456 12s 152ms/step - loss: 0.1225 - acc: 0.9870 - val_loss: 0.4545 - val_acc: 0.9074 457 Epoch 2728/5000 458 12s 152ms/step - loss: 0.1225 - acc: 0.9873 - val_loss: 0.4643 - val_acc: 0.9047 459 Epoch 2729/5000 460 12s 152ms/step - loss: 0.1240 - acc: 0.9872 - val_loss: 0.4518 - val_acc: 0.9052 461 Epoch 2730/5000 462 12s 152ms/step - loss: 0.1248 - acc: 0.9867 - val_loss: 0.4580 - val_acc: 0.9045 463 Epoch 2731/5000 464 12s 152ms/step - loss: 0.1258 - acc: 0.9863 - val_loss: 0.4620 - val_acc: 0.9028 465 Epoch 2732/5000 466 12s 152ms/step - loss: 0.1273 - acc: 0.9862 - val_loss: 0.4536 - val_acc: 0.9053 467 Epoch 2733/5000 468 12s 152ms/step - loss: 0.1273 - acc: 0.9862 - val_loss: 0.4440 - val_acc: 0.9074 469 Epoch 2734/5000 470 12s 152ms/step - loss: 0.1257 - acc: 0.9856 - val_loss: 0.4456 - val_acc: 0.9043 471 Epoch 2735/5000 472 12s 151ms/step - loss: 0.1231 - acc: 0.9876 - val_loss: 0.4559 - val_acc: 0.9051 473 Epoch 2736/5000 474 12s 152ms/step - loss: 0.1254 - acc: 0.9858 - val_loss: 0.4470 - val_acc: 0.9077 475 Epoch 2737/5000 476 12s 152ms/step - loss: 0.1246 - acc: 0.9866 - val_loss: 0.4549 - val_acc: 0.9048 477 Epoch 2738/5000 478 12s 152ms/step - loss: 0.1223 - acc: 0.9874 - val_loss: 0.4676 - val_acc: 0.9047 479 Epoch 2739/5000 480 12s 151ms/step - loss: 0.1228 - acc: 0.9871 - val_loss: 0.4466 - val_acc: 0.9072 481 Epoch 2740/5000 482 12s 152ms/step - loss: 0.1236 - acc: 0.9869 - val_loss: 0.4514 - val_acc: 0.9045 483 Epoch 2741/5000 484 12s 151ms/step - loss: 0.1271 - acc: 0.9853 - val_loss: 0.4638 - val_acc: 0.9020 485 Epoch 2742/5000 486 12s 152ms/step - loss: 0.1256 - acc: 0.9860 - val_loss: 0.4513 - val_acc: 0.9084 487 Epoch 2743/5000 488 12s 152ms/step - loss: 0.1241 - acc: 0.9868 - val_loss: 0.4537 - val_acc: 0.9090 489 Epoch 2744/5000 490 12s 152ms/step - loss: 0.1242 - acc: 0.9866 - val_loss: 0.4572 - val_acc: 0.9058 491 Epoch 2745/5000 492 12s 152ms/step - loss: 0.1251 - acc: 0.9858 - val_loss: 0.4705 - val_acc: 0.9030 493 Epoch 2746/5000 494 12s 151ms/step - loss: 0.1258 - acc: 0.9858 - val_loss: 0.4691 - val_acc: 0.9034 495 Epoch 2747/5000 496 12s 151ms/step - loss: 0.1255 - acc: 0.9865 - val_loss: 0.4597 - val_acc: 0.9013 497 Epoch 2748/5000 498 12s 151ms/step - loss: 0.1255 - acc: 0.9862 - val_loss: 0.4440 - val_acc: 0.9070 499 Epoch 2749/5000 500 12s 152ms/step - loss: 0.1256 - acc: 0.9856 - val_loss: 0.4690 - val_acc: 0.9029 501 Epoch 2750/5000 502 12s 152ms/step - loss: 0.1253 - acc: 0.9864 - val_loss: 0.4515 - val_acc: 0.9037 503 Epoch 2751/5000 504 12s 151ms/step - loss: 0.1230 - acc: 0.9869 - val_loss: 0.4741 - val_acc: 0.9035 505 Epoch 2752/5000 506 12s 151ms/step - loss: 0.1289 - acc: 0.9846 - val_loss: 0.4739 - val_acc: 0.9010 507 Epoch 2753/5000 508 12s 151ms/step - loss: 0.1281 - acc: 0.9854 - val_loss: 0.4494 - val_acc: 0.9033 509 Epoch 2754/5000 510 12s 151ms/step - loss: 0.1258 - acc: 0.9863 - val_loss: 0.4558 - val_acc: 0.9058 511 Epoch 2755/5000 512 12s 151ms/step - loss: 0.1261 - acc: 0.9859 - val_loss: 0.4617 - val_acc: 0.9045 513 Epoch 2756/5000 514 12s 151ms/step - loss: 0.1250 - acc: 0.9864 - val_loss: 0.4554 - val_acc: 0.9052 515 Epoch 2757/5000 516 12s 151ms/step - loss: 0.1262 - acc: 0.9859 - val_loss: 0.4478 - val_acc: 0.9060 517 Epoch 2758/5000 518 12s 152ms/step - loss: 0.1225 - acc: 0.9872 - val_loss: 0.4455 - val_acc: 0.9047 519 Epoch 2759/5000 520 12s 151ms/step - loss: 0.1234 - acc: 0.9868 - val_loss: 0.4477 - val_acc: 0.9067 521 Epoch 2760/5000 522 12s 152ms/step - loss: 0.1271 - acc: 0.9859 - val_loss: 0.4535 - val_acc: 0.9032 523 Epoch 2761/5000 524 12s 151ms/step - loss: 0.1238 - acc: 0.9867 - val_loss: 0.4691 - val_acc: 0.9012 525 Epoch 2762/5000 526 12s 151ms/step - loss: 0.1233 - acc: 0.9868 - val_loss: 0.4584 - val_acc: 0.9029 527 Epoch 2763/5000 528 12s 151ms/step - loss: 0.1244 - acc: 0.9871 - val_loss: 0.4508 - val_acc: 0.9016 529 Epoch 2764/5000 530 12s 151ms/step - loss: 0.1233 - acc: 0.9865 - val_loss: 0.4672 - val_acc: 0.9027 531 Epoch 2765/5000 532 12s 151ms/step - loss: 0.1264 - acc: 0.9863 - val_loss: 0.4467 - val_acc: 0.9066 533 Epoch 2766/5000 534 12s 151ms/step - loss: 0.1244 - acc: 0.9866 - val_loss: 0.4622 - val_acc: 0.9018 535 Epoch 2767/5000 536 12s 151ms/step - loss: 0.1231 - acc: 0.9872 - val_loss: 0.4463 - val_acc: 0.9054 537 Epoch 2768/5000 538 12s 151ms/step - loss: 0.1241 - acc: 0.9867 - val_loss: 0.4526 - val_acc: 0.9055 539 Epoch 2769/5000 540 12s 151ms/step - loss: 0.1271 - acc: 0.9854 - val_loss: 0.4525 - val_acc: 0.9023 541 Epoch 2770/5000 542 12s 152ms/step - loss: 0.1225 - acc: 0.9878 - val_loss: 0.4537 - val_acc: 0.9046 543 Epoch 2771/5000 544 12s 151ms/step - loss: 0.1248 - acc: 0.9865 - val_loss: 0.4617 - val_acc: 0.9050 545 Epoch 2772/5000 546 12s 151ms/step - loss: 0.1251 - acc: 0.9861 - val_loss: 0.4598 - val_acc: 0.9050 547 Epoch 2773/5000 548 12s 151ms/step - loss: 0.1268 - acc: 0.9859 - val_loss: 0.4630 - val_acc: 0.9044 549 Epoch 2774/5000 550 12s 152ms/step - loss: 0.1231 - acc: 0.9874 - val_loss: 0.4568 - val_acc: 0.9015 551 Epoch 2775/5000 552 12s 152ms/step - loss: 0.1254 - acc: 0.9861 - val_loss: 0.4578 - val_acc: 0.9038 553 Epoch 2776/5000 554 12s 151ms/step - loss: 0.1225 - acc: 0.9873 - val_loss: 0.4647 - val_acc: 0.9030 555 Epoch 2777/5000 556 12s 151ms/step - loss: 0.1227 - acc: 0.9874 - val_loss: 0.4515 - val_acc: 0.9047 557 Epoch 2778/5000 558 12s 151ms/step - loss: 0.1261 - acc: 0.9858 - val_loss: 0.4580 - val_acc: 0.9032 559 Epoch 2779/5000 560 12s 151ms/step - loss: 0.1240 - acc: 0.9866 - val_loss: 0.4722 - val_acc: 0.9035 561 Epoch 2780/5000 562 12s 151ms/step - loss: 0.1236 - acc: 0.9871 - val_loss: 0.4720 - val_acc: 0.9014 563 Epoch 2781/5000 564 12s 151ms/step - loss: 0.1264 - acc: 0.9859 - val_loss: 0.4523 - val_acc: 0.9031 565 Epoch 2782/5000 566 12s 152ms/step - loss: 0.1261 - acc: 0.9859 - val_loss: 0.4556 - val_acc: 0.9046 567 Epoch 2783/5000 568 12s 151ms/step - loss: 0.1266 - acc: 0.9859 - val_loss: 0.4390 - val_acc: 0.9088 569 Epoch 2784/5000 570 12s 151ms/step - loss: 0.1244 - acc: 0.9862 - val_loss: 0.4533 - val_acc: 0.9033 571 Epoch 2785/5000 572 12s 152ms/step - loss: 0.1227 - acc: 0.9871 - val_loss: 0.4548 - val_acc: 0.9038 573 Epoch 2786/5000 574 12s 152ms/step - loss: 0.1229 - acc: 0.9870 - val_loss: 0.4468 - val_acc: 0.9066 575 Epoch 2787/5000 576 12s 152ms/step - loss: 0.1220 - acc: 0.9870 - val_loss: 0.4466 - val_acc: 0.9060 577 Epoch 2788/5000 578 12s 152ms/step - loss: 0.1294 - acc: 0.9849 - val_loss: 0.4455 - val_acc: 0.9025 579 Epoch 2789/5000 580 12s 151ms/step - loss: 0.1251 - acc: 0.9866 - val_loss: 0.4618 - val_acc: 0.9024 581 Epoch 2790/5000 582 12s 151ms/step - loss: 0.1235 - acc: 0.9871 - val_loss: 0.4551 - val_acc: 0.9035 583 Epoch 2791/5000 584 12s 152ms/step - loss: 0.1260 - acc: 0.9858 - val_loss: 0.4594 - val_acc: 0.9025 585 Epoch 2792/5000 586 12s 151ms/step - loss: 0.1207 - acc: 0.9879 - val_loss: 0.4490 - val_acc: 0.9024 587 Epoch 2793/5000 588 12s 151ms/step - loss: 0.1224 - acc: 0.9874 - val_loss: 0.4498 - val_acc: 0.9037 589 Epoch 2794/5000 590 12s 151ms/step - loss: 0.1236 - acc: 0.9874 - val_loss: 0.4533 - val_acc: 0.9008 591 Epoch 2795/5000 592 12s 152ms/step - loss: 0.1224 - acc: 0.9874 - val_loss: 0.4410 - val_acc: 0.9074 593 Epoch 2796/5000 594 12s 152ms/step - loss: 0.1237 - acc: 0.9867 - val_loss: 0.4506 - val_acc: 0.9057 595 Epoch 2797/5000 596 12s 151ms/step - loss: 0.1237 - acc: 0.9868 - val_loss: 0.4451 - val_acc: 0.9042 597 Epoch 2798/5000 598 12s 151ms/step - loss: 0.1247 - acc: 0.9862 - val_loss: 0.4678 - val_acc: 0.9009 599 Epoch 2799/5000 600 12s 151ms/step - loss: 0.1262 - acc: 0.9864 - val_loss: 0.4639 - val_acc: 0.9024 601 Epoch 2800/5000 602 12s 151ms/step - loss: 0.1269 - acc: 0.9857 - val_loss: 0.4550 - val_acc: 0.9029 603 Epoch 2801/5000 604 12s 151ms/step - loss: 0.1273 - acc: 0.9857 - val_loss: 0.4514 - val_acc: 0.9036 605 Epoch 2802/5000 606 12s 152ms/step - loss: 0.1240 - acc: 0.9869 - val_loss: 0.4525 - val_acc: 0.9031 607 Epoch 2803/5000 608 12s 151ms/step - loss: 0.1244 - acc: 0.9865 - val_loss: 0.4652 - val_acc: 0.9018 609 Epoch 2804/5000 610 12s 151ms/step - loss: 0.1265 - acc: 0.9864 - val_loss: 0.4765 - val_acc: 0.8992 611 Epoch 2805/5000 612 12s 151ms/step - loss: 0.1260 - acc: 0.9855 - val_loss: 0.4589 - val_acc: 0.9025 613 Epoch 2806/5000 614 12s 151ms/step - loss: 0.1244 - acc: 0.9870 - val_loss: 0.4605 - val_acc: 0.9039 615 Epoch 2807/5000 616 12s 151ms/step - loss: 0.1243 - acc: 0.9864 - val_loss: 0.4580 - val_acc: 0.9028 617 Epoch 2808/5000 618 12s 152ms/step - loss: 0.1213 - acc: 0.9874 - val_loss: 0.4514 - val_acc: 0.9060 619 Epoch 2809/5000 620 12s 151ms/step - loss: 0.1213 - acc: 0.9876 - val_loss: 0.4663 - val_acc: 0.9008 621 Epoch 2810/5000 622 12s 152ms/step - loss: 0.1249 - acc: 0.9870 - val_loss: 0.4634 - val_acc: 0.9025 623 Epoch 2811/5000 624 12s 151ms/step - loss: 0.1252 - acc: 0.9865 - val_loss: 0.4576 - val_acc: 0.9057 625 Epoch 2812/5000 626 12s 151ms/step - loss: 0.1250 - acc: 0.9861 - val_loss: 0.4713 - val_acc: 0.9003 627 Epoch 2813/5000 628 12s 151ms/step - loss: 0.1257 - acc: 0.9859 - val_loss: 0.4511 - val_acc: 0.9059 629 Epoch 2814/5000 630 12s 152ms/step - loss: 0.1257 - acc: 0.9867 - val_loss: 0.4700 - val_acc: 0.9009 631 Epoch 2815/5000 632 12s 151ms/step - loss: 0.1253 - acc: 0.9860 - val_loss: 0.4602 - val_acc: 0.9046 633 Epoch 2816/5000 634 12s 151ms/step - loss: 0.1262 - acc: 0.9856 - val_loss: 0.4570 - val_acc: 0.9012 635 Epoch 2817/5000 636 12s 151ms/step - loss: 0.1256 - acc: 0.9861 - val_loss: 0.4609 - val_acc: 0.9020 637 Epoch 2818/5000 638 12s 151ms/step - loss: 0.1262 - acc: 0.9862 - val_loss: 0.4482 - val_acc: 0.9059 639 Epoch 2819/5000 640 12s 151ms/step - loss: 0.1249 - acc: 0.9865 - val_loss: 0.4531 - val_acc: 0.9058 641 Epoch 2820/5000 642 12s 151ms/step - loss: 0.1225 - acc: 0.9876 - val_loss: 0.4457 - val_acc: 0.9053 643 Epoch 2821/5000 644 12s 151ms/step - loss: 0.1226 - acc: 0.9871 - val_loss: 0.4470 - val_acc: 0.9061 645 Epoch 2822/5000 646 12s 152ms/step - loss: 0.1253 - acc: 0.9856 - val_loss: 0.4415 - val_acc: 0.9091 647 Epoch 2823/5000 648 12s 151ms/step - loss: 0.1273 - acc: 0.9849 - val_loss: 0.4557 - val_acc: 0.9026 649 Epoch 2824/5000 650 12s 151ms/step - loss: 0.1238 - acc: 0.9873 - val_loss: 0.4350 - val_acc: 0.9062 651 Epoch 2825/5000 652 12s 151ms/step - loss: 0.1216 - acc: 0.9875 - val_loss: 0.4519 - val_acc: 0.9055 653 Epoch 2826/5000 654 12s 151ms/step - loss: 0.1245 - acc: 0.9867 - val_loss: 0.4502 - val_acc: 0.9055 655 Epoch 2827/5000 656 12s 151ms/step - loss: 0.1230 - acc: 0.9872 - val_loss: 0.4619 - val_acc: 0.9049 657 Epoch 2828/5000 658 12s 151ms/step - loss: 0.1238 - acc: 0.9869 - val_loss: 0.4563 - val_acc: 0.9032 659 Epoch 2829/5000 660 12s 152ms/step - loss: 0.1243 - acc: 0.9863 - val_loss: 0.4650 - val_acc: 0.9017 661 Epoch 2830/5000 662 12s 152ms/step - loss: 0.1241 - acc: 0.9869 - val_loss: 0.4628 - val_acc: 0.9023 663 Epoch 2831/5000 664 12s 151ms/step - loss: 0.1268 - acc: 0.9857 - val_loss: 0.4599 - val_acc: 0.9058 665 Epoch 2832/5000 666 12s 151ms/step - loss: 0.1234 - acc: 0.9871 - val_loss: 0.4551 - val_acc: 0.9061 667 Epoch 2833/5000 668 12s 151ms/step - loss: 0.1235 - acc: 0.9865 - val_loss: 0.4608 - val_acc: 0.9055 669 Epoch 2834/5000 670 12s 151ms/step - loss: 0.1257 - acc: 0.9866 - val_loss: 0.4463 - val_acc: 0.9076 671 Epoch 2835/5000 672 12s 151ms/step - loss: 0.1231 - acc: 0.9869 - val_loss: 0.4648 - val_acc: 0.8993 673 Epoch 2836/5000 674 12s 151ms/step - loss: 0.1246 - acc: 0.9864 - val_loss: 0.4587 - val_acc: 0.9045 675 Epoch 2837/5000 676 12s 152ms/step - loss: 0.1254 - acc: 0.9865 - val_loss: 0.4570 - val_acc: 0.9009 677 Epoch 2838/5000 678 12s 151ms/step - loss: 0.1257 - acc: 0.9861 - val_loss: 0.4606 - val_acc: 0.9026 679 Epoch 2839/5000 680 12s 152ms/step - loss: 0.1248 - acc: 0.9864 - val_loss: 0.4673 - val_acc: 0.9034 681 Epoch 2840/5000 682 12s 151ms/step - loss: 0.1253 - acc: 0.9862 - val_loss: 0.4600 - val_acc: 0.9042 683 Epoch 2841/5000 684 12s 151ms/step - loss: 0.1243 - acc: 0.9866 - val_loss: 0.4696 - val_acc: 0.9013 685 Epoch 2842/5000 686 12s 150ms/step - loss: 0.1240 - acc: 0.9871 - val_loss: 0.4504 - val_acc: 0.9052 687 Epoch 2843/5000 688 12s 151ms/step - loss: 0.1238 - acc: 0.9865 - val_loss: 0.4590 - val_acc: 0.9025 689 Epoch 2844/5000 690 12s 151ms/step - loss: 0.1246 - acc: 0.9866 - val_loss: 0.4587 - val_acc: 0.9003 691 Epoch 2845/5000 692 12s 151ms/step - loss: 0.1252 - acc: 0.9865 - val_loss: 0.4593 - val_acc: 0.9022 693 Epoch 2846/5000 694 12s 152ms/step - loss: 0.1225 - acc: 0.9876 - val_loss: 0.4584 - val_acc: 0.9064 695 Epoch 2847/5000 696 12s 151ms/step - loss: 0.1250 - acc: 0.9866 - val_loss: 0.4614 - val_acc: 0.9063 697 Epoch 2848/5000 698 12s 152ms/step - loss: 0.1252 - acc: 0.9864 - val_loss: 0.4774 - val_acc: 0.9039 699 Epoch 2849/5000 700 12s 152ms/step - loss: 0.1243 - acc: 0.9868 - val_loss: 0.4544 - val_acc: 0.9066 701 Epoch 2850/5000 702 12s 152ms/step - loss: 0.1255 - acc: 0.9861 - val_loss: 0.4497 - val_acc: 0.9040 703 Epoch 2851/5000 704 12s 151ms/step - loss: 0.1236 - acc: 0.9867 - val_loss: 0.4512 - val_acc: 0.9018 705 Epoch 2852/5000 706 12s 152ms/step - loss: 0.1258 - acc: 0.9860 - val_loss: 0.4568 - val_acc: 0.9042 707 Epoch 2853/5000 708 12s 151ms/step - loss: 0.1212 - acc: 0.9871 - val_loss: 0.4575 - val_acc: 0.9030 709 Epoch 2854/5000 710 12s 151ms/step - loss: 0.1240 - acc: 0.9865 - val_loss: 0.4592 - val_acc: 0.9024 711 Epoch 2855/5000 712 12s 151ms/step - loss: 0.1231 - acc: 0.9869 - val_loss: 0.4464 - val_acc: 0.9079 713 Epoch 2856/5000 714 12s 152ms/step - loss: 0.1229 - acc: 0.9872 - val_loss: 0.4571 - val_acc: 0.9039 715 Epoch 2857/5000 716 12s 152ms/step - loss: 0.1237 - acc: 0.9871 - val_loss: 0.4527 - val_acc: 0.9056 717 Epoch 2858/5000 718 12s 152ms/step - loss: 0.1224 - acc: 0.9872 - val_loss: 0.4403 - val_acc: 0.9081 719 Epoch 2859/5000 720 12s 151ms/step - loss: 0.1249 - acc: 0.9859 - val_loss: 0.4666 - val_acc: 0.9017 721 Epoch 2860/5000 722 12s 151ms/step - loss: 0.1259 - acc: 0.9859 - val_loss: 0.4420 - val_acc: 0.9056 723 Epoch 2861/5000 724 12s 151ms/step - loss: 0.1240 - acc: 0.9865 - val_loss: 0.4547 - val_acc: 0.9039 725 Epoch 2862/5000 726 12s 151ms/step - loss: 0.1240 - acc: 0.9868 - val_loss: 0.4583 - val_acc: 0.9038 727 Epoch 2863/5000 728 12s 151ms/step - loss: 0.1255 - acc: 0.9856 - val_loss: 0.4665 - val_acc: 0.9044 729 Epoch 2864/5000 730 12s 152ms/step - loss: 0.1264 - acc: 0.9862 - val_loss: 0.4568 - val_acc: 0.9049 731 Epoch 2865/5000 732 12s 151ms/step - loss: 0.1278 - acc: 0.9852 - val_loss: 0.4730 - val_acc: 0.9007 733 Epoch 2866/5000 734 12s 152ms/step - loss: 0.1257 - acc: 0.9861 - val_loss: 0.4602 - val_acc: 0.9011 735 Epoch 2867/5000 736 12s 152ms/step - loss: 0.1275 - acc: 0.9862 - val_loss: 0.4459 - val_acc: 0.9055 737 Epoch 2868/5000 738 12s 152ms/step - loss: 0.1265 - acc: 0.9858 - val_loss: 0.4441 - val_acc: 0.9048 739 Epoch 2869/5000 740 12s 151ms/step - loss: 0.1243 - acc: 0.9874 - val_loss: 0.4566 - val_acc: 0.9034 741 Epoch 2870/5000 742 12s 151ms/step - loss: 0.1261 - acc: 0.9864 - val_loss: 0.4653 - val_acc: 0.9012 743 Epoch 2871/5000 744 12s 152ms/step - loss: 0.1267 - acc: 0.9860 - val_loss: 0.4621 - val_acc: 0.8996 745 Epoch 2872/5000 746 12s 151ms/step - loss: 0.1240 - acc: 0.9863 - val_loss: 0.4517 - val_acc: 0.9050 747 Epoch 2873/5000 748 12s 151ms/step - loss: 0.1240 - acc: 0.9867 - val_loss: 0.4478 - val_acc: 0.9058 749 Epoch 2874/5000 750 12s 151ms/step - loss: 0.1241 - acc: 0.9865 - val_loss: 0.4507 - val_acc: 0.9058 751 Epoch 2875/5000 752 12s 152ms/step - loss: 0.1218 - acc: 0.9879 - val_loss: 0.4455 - val_acc: 0.9055 753 Epoch 2876/5000 754 12s 151ms/step - loss: 0.1252 - acc: 0.9859 - val_loss: 0.4639 - val_acc: 0.9012 755 Epoch 2877/5000 756 12s 151ms/step - loss: 0.1217 - acc: 0.9872 - val_loss: 0.4713 - val_acc: 0.9009 757 Epoch 2878/5000 758 12s 152ms/step - loss: 0.1227 - acc: 0.9877 - val_loss: 0.4590 - val_acc: 0.9031 759 Epoch 2879/5000 760 12s 150ms/step - loss: 0.1247 - acc: 0.9863 - val_loss: 0.4390 - val_acc: 0.9054 761 Epoch 2880/5000 762 12s 152ms/step - loss: 0.1251 - acc: 0.9865 - val_loss: 0.4582 - val_acc: 0.9025 763 Epoch 2881/5000 764 12s 151ms/step - loss: 0.1261 - acc: 0.9859 - val_loss: 0.4480 - val_acc: 0.9053 765 Epoch 2882/5000 766 12s 152ms/step - loss: 0.1228 - acc: 0.9869 - val_loss: 0.4430 - val_acc: 0.9089 767 Epoch 2883/5000 768 12s 151ms/step - loss: 0.1215 - acc: 0.9869 - val_loss: 0.4476 - val_acc: 0.9061 769 Epoch 2884/5000 770 12s 151ms/step - loss: 0.1272 - acc: 0.9856 - val_loss: 0.4586 - val_acc: 0.9062 771 Epoch 2885/5000 772 12s 151ms/step - loss: 0.1243 - acc: 0.9861 - val_loss: 0.4557 - val_acc: 0.9021 773 Epoch 2886/5000 774 12s 151ms/step - loss: 0.1240 - acc: 0.9858 - val_loss: 0.4631 - val_acc: 0.9041 775 Epoch 2887/5000 776 12s 152ms/step - loss: 0.1234 - acc: 0.9870 - val_loss: 0.4378 - val_acc: 0.9074 777 Epoch 2888/5000 778 12s 152ms/step - loss: 0.1245 - acc: 0.9865 - val_loss: 0.4415 - val_acc: 0.9072 779 Epoch 2889/5000 780 12s 151ms/step - loss: 0.1252 - acc: 0.9865 - val_loss: 0.4535 - val_acc: 0.9072 781 Epoch 2890/5000 782 12s 151ms/step - loss: 0.1218 - acc: 0.9869 - val_loss: 0.4449 - val_acc: 0.9073 783 Epoch 2891/5000 784 12s 152ms/step - loss: 0.1268 - acc: 0.9852 - val_loss: 0.4485 - val_acc: 0.9015 785 Epoch 2892/5000 786 12s 152ms/step - loss: 0.1248 - acc: 0.9865 - val_loss: 0.4578 - val_acc: 0.9034 787 Epoch 2893/5000 788 12s 151ms/step - loss: 0.1236 - acc: 0.9870 - val_loss: 0.4452 - val_acc: 0.9067 789 Epoch 2894/5000 790 12s 151ms/step - loss: 0.1251 - acc: 0.9866 - val_loss: 0.4537 - val_acc: 0.9031 791 Epoch 2895/5000 792 12s 151ms/step - loss: 0.1268 - acc: 0.9859 - val_loss: 0.4650 - val_acc: 0.9049 793 Epoch 2896/5000 794 12s 151ms/step - loss: 0.1245 - acc: 0.9864 - val_loss: 0.4558 - val_acc: 0.9049 795 Epoch 2897/5000 796 12s 152ms/step - loss: 0.1225 - acc: 0.9867 - val_loss: 0.4567 - val_acc: 0.9062 797 Epoch 2898/5000 798 12s 151ms/step - loss: 0.1227 - acc: 0.9873 - val_loss: 0.4518 - val_acc: 0.9032 799 Epoch 2899/5000 800 12s 151ms/step - loss: 0.1224 - acc: 0.9877 - val_loss: 0.4370 - val_acc: 0.9065 801 Epoch 2900/5000 802 12s 151ms/step - loss: 0.1232 - acc: 0.9870 - val_loss: 0.4514 - val_acc: 0.9074 803 Epoch 2901/5000 804 12s 151ms/step - loss: 0.1230 - acc: 0.9867 - val_loss: 0.4497 - val_acc: 0.9035 805 Epoch 2902/5000 806 12s 152ms/step - loss: 0.1242 - acc: 0.9870 - val_loss: 0.4468 - val_acc: 0.9046 807 Epoch 2903/5000 808 12s 151ms/step - loss: 0.1241 - acc: 0.9866 - val_loss: 0.4631 - val_acc: 0.9030 809 Epoch 2904/5000 810 12s 152ms/step - loss: 0.1232 - acc: 0.9871 - val_loss: 0.4451 - val_acc: 0.9057 811 Epoch 2905/5000 812 12s 152ms/step - loss: 0.1217 - acc: 0.9869 - val_loss: 0.4466 - val_acc: 0.9079 813 Epoch 2906/5000 814 12s 150ms/step - loss: 0.1249 - acc: 0.9861 - val_loss: 0.4484 - val_acc: 0.9049 815 Epoch 2907/5000 816 12s 152ms/step - loss: 0.1253 - acc: 0.9865 - val_loss: 0.4467 - val_acc: 0.9079 817 Epoch 2908/5000 818 12s 152ms/step - loss: 0.1250 - acc: 0.9866 - val_loss: 0.4523 - val_acc: 0.9040 819 Epoch 2909/5000 820 12s 151ms/step - loss: 0.1261 - acc: 0.9855 - val_loss: 0.4477 - val_acc: 0.9075 821 Epoch 2910/5000 822 12s 152ms/step - loss: 0.1247 - acc: 0.9866 - val_loss: 0.4358 - val_acc: 0.9090 823 Epoch 2911/5000 824 12s 150ms/step - loss: 0.1222 - acc: 0.9876 - val_loss: 0.4622 - val_acc: 0.9016 825 Epoch 2912/5000 826 12s 152ms/step - loss: 0.1246 - acc: 0.9863 - val_loss: 0.4487 - val_acc: 0.9069 827 Epoch 2913/5000 828 12s 152ms/step - loss: 0.1231 - acc: 0.9873 - val_loss: 0.4466 - val_acc: 0.9051 829 Epoch 2914/5000 830 12s 151ms/step - loss: 0.1240 - acc: 0.9866 - val_loss: 0.4554 - val_acc: 0.9051 831 Epoch 2915/5000 832 12s 152ms/step - loss: 0.1232 - acc: 0.9872 - val_loss: 0.4558 - val_acc: 0.9063 833 Epoch 2916/5000 834 12s 151ms/step - loss: 0.1217 - acc: 0.9870 - val_loss: 0.4533 - val_acc: 0.9055 835 Epoch 2917/5000 836 12s 151ms/step - loss: 0.1242 - acc: 0.9866 - val_loss: 0.4584 - val_acc: 0.9017 837 Epoch 2918/5000 838 12s 152ms/step - loss: 0.1233 - acc: 0.9871 - val_loss: 0.4594 - val_acc: 0.9025 839 Epoch 2919/5000 840 12s 151ms/step - loss: 0.1213 - acc: 0.9876 - val_loss: 0.4582 - val_acc: 0.9034 841 Epoch 2920/5000 842 12s 151ms/step - loss: 0.1226 - acc: 0.9872 - val_loss: 0.4494 - val_acc: 0.9047 843 Epoch 2921/5000 844 12s 151ms/step - loss: 0.1218 - acc: 0.9872 - val_loss: 0.4576 - val_acc: 0.9066 845 Epoch 2922/5000 846 12s 151ms/step - loss: 0.1243 - acc: 0.9863 - val_loss: 0.4597 - val_acc: 0.9055 847 Epoch 2923/5000 848 12s 152ms/step - loss: 0.1269 - acc: 0.9852 - val_loss: 0.4563 - val_acc: 0.9054 849 Epoch 2924/5000 850 12s 151ms/step - loss: 0.1242 - acc: 0.9865 - val_loss: 0.4465 - val_acc: 0.9038 851 Epoch 2925/5000 852 12s 152ms/step - loss: 0.1218 - acc: 0.9872 - val_loss: 0.4531 - val_acc: 0.9027 853 Epoch 2926/5000 854 12s 151ms/step - loss: 0.1247 - acc: 0.9863 - val_loss: 0.4551 - val_acc: 0.9046 855 Epoch 2927/5000 856 12s 151ms/step - loss: 0.1242 - acc: 0.9866 - val_loss: 0.4591 - val_acc: 0.9019 857 Epoch 2928/5000 858 12s 151ms/step - loss: 0.1232 - acc: 0.9867 - val_loss: 0.4550 - val_acc: 0.9037 859 Epoch 2929/5000 860 12s 151ms/step - loss: 0.1216 - acc: 0.9879 - val_loss: 0.4495 - val_acc: 0.9054 861 Epoch 2930/5000 862 12s 152ms/step - loss: 0.1228 - acc: 0.9871 - val_loss: 0.4478 - val_acc: 0.9043 863 Epoch 2931/5000 864 12s 152ms/step - loss: 0.1243 - acc: 0.9859 - val_loss: 0.4601 - val_acc: 0.9025 865 Epoch 2932/5000 866 12s 152ms/step - loss: 0.1238 - acc: 0.9865 - val_loss: 0.4561 - val_acc: 0.9050 867 Epoch 2933/5000 868 12s 152ms/step - loss: 0.1233 - acc: 0.9873 - val_loss: 0.4625 - val_acc: 0.9024 869 Epoch 2934/5000 870 12s 152ms/step - loss: 0.1245 - acc: 0.9859 - val_loss: 0.4558 - val_acc: 0.9025 871 Epoch 2935/5000 872 12s 152ms/step - loss: 0.1252 - acc: 0.9862 - val_loss: 0.4648 - val_acc: 0.9030 873 Epoch 2936/5000 874 12s 151ms/step - loss: 0.1229 - acc: 0.9872 - val_loss: 0.4648 - val_acc: 0.9024 875 Epoch 2937/5000 876 12s 152ms/step - loss: 0.1233 - acc: 0.9867 - val_loss: 0.4577 - val_acc: 0.9015 877 Epoch 2938/5000 878 12s 151ms/step - loss: 0.1266 - acc: 0.9854 - val_loss: 0.4721 - val_acc: 0.9011 879 Epoch 2939/5000 880 12s 151ms/step - loss: 0.1236 - acc: 0.9868 - val_loss: 0.4562 - val_acc: 0.9050 881 Epoch 2940/5000 882 12s 152ms/step - loss: 0.1221 - acc: 0.9868 - val_loss: 0.4583 - val_acc: 0.9046 883 Epoch 2941/5000 884 12s 152ms/step - loss: 0.1244 - acc: 0.9863 - val_loss: 0.4601 - val_acc: 0.9039 885 Epoch 2942/5000 886 12s 151ms/step - loss: 0.1234 - acc: 0.9873 - val_loss: 0.4710 - val_acc: 0.9021 887 Epoch 2943/5000 888 12s 152ms/step - loss: 0.1227 - acc: 0.9869 - val_loss: 0.4574 - val_acc: 0.9057 889 Epoch 2944/5000 890 12s 151ms/step - loss: 0.1245 - acc: 0.9862 - val_loss: 0.4752 - val_acc: 0.9023 891 Epoch 2945/5000 892 12s 151ms/step - loss: 0.1228 - acc: 0.9866 - val_loss: 0.4870 - val_acc: 0.8995 893 Epoch 2946/5000 894 12s 151ms/step - loss: 0.1230 - acc: 0.9870 - val_loss: 0.4680 - val_acc: 0.9030 895 Epoch 2947/5000 896 12s 151ms/step - loss: 0.1240 - acc: 0.9870 - val_loss: 0.4631 - val_acc: 0.9043 897 Epoch 2948/5000 898 12s 151ms/step - loss: 0.1244 - acc: 0.9863 - val_loss: 0.4531 - val_acc: 0.9046 899 Epoch 2949/5000 900 12s 151ms/step - loss: 0.1265 - acc: 0.9857 - val_loss: 0.4524 - val_acc: 0.9047 901 Epoch 2950/5000 902 12s 151ms/step - loss: 0.1242 - acc: 0.9870 - val_loss: 0.4612 - val_acc: 0.9021 903 Epoch 2951/5000 904 12s 152ms/step - loss: 0.1252 - acc: 0.9861 - val_loss: 0.4594 - val_acc: 0.9056 905 Epoch 2952/5000 906 12s 151ms/step - loss: 0.1238 - acc: 0.9865 - val_loss: 0.4678 - val_acc: 0.9039 907 Epoch 2953/5000 908 12s 151ms/step - loss: 0.1195 - acc: 0.9883 - val_loss: 0.4595 - val_acc: 0.9059 909 Epoch 2954/5000 910 12s 151ms/step - loss: 0.1219 - acc: 0.9870 - val_loss: 0.4533 - val_acc: 0.9056 911 Epoch 2955/5000 912 12s 152ms/step - loss: 0.1266 - acc: 0.9854 - val_loss: 0.4631 - val_acc: 0.9023 913 Epoch 2956/5000 914 12s 151ms/step - loss: 0.1252 - acc: 0.9856 - val_loss: 0.4567 - val_acc: 0.9050 915 Epoch 2957/5000 916 12s 151ms/step - loss: 0.1263 - acc: 0.9862 - val_loss: 0.4424 - val_acc: 0.9104 917 Epoch 2958/5000 918 12s 151ms/step - loss: 0.1221 - acc: 0.9871 - val_loss: 0.4534 - val_acc: 0.9059 919 Epoch 2959/5000 920 12s 152ms/step - loss: 0.1227 - acc: 0.9869 - val_loss: 0.4523 - val_acc: 0.9097 921 Epoch 2960/5000 922 12s 150ms/step - loss: 0.1237 - acc: 0.9874 - val_loss: 0.4554 - val_acc: 0.9057 923 Epoch 2961/5000 924 12s 151ms/step - loss: 0.1246 - acc: 0.9860 - val_loss: 0.4488 - val_acc: 0.9077 925 Epoch 2962/5000 926 12s 152ms/step - loss: 0.1235 - acc: 0.9872 - val_loss: 0.4559 - val_acc: 0.9021 927 Epoch 2963/5000 928 12s 151ms/step - loss: 0.1226 - acc: 0.9873 - val_loss: 0.4650 - val_acc: 0.9019 929 Epoch 2964/5000 930 12s 151ms/step - loss: 0.1259 - acc: 0.9858 - val_loss: 0.4653 - val_acc: 0.9009 931 Epoch 2965/5000 932 12s 151ms/step - loss: 0.1259 - acc: 0.9861 - val_loss: 0.4566 - val_acc: 0.9026 933 Epoch 2966/5000 934 12s 151ms/step - loss: 0.1221 - acc: 0.9873 - val_loss: 0.4626 - val_acc: 0.9038 935 Epoch 2967/5000 936 12s 152ms/step - loss: 0.1251 - acc: 0.9861 - val_loss: 0.4591 - val_acc: 0.9049 937 Epoch 2968/5000 938 12s 151ms/step - loss: 0.1242 - acc: 0.9861 - val_loss: 0.4526 - val_acc: 0.9073 939 Epoch 2969/5000 940 12s 152ms/step - loss: 0.1235 - acc: 0.9865 - val_loss: 0.4467 - val_acc: 0.9064 941 Epoch 2970/5000 942 12s 151ms/step - loss: 0.1272 - acc: 0.9854 - val_loss: 0.4561 - val_acc: 0.9058 943 Epoch 2971/5000 944 12s 152ms/step - loss: 0.1233 - acc: 0.9869 - val_loss: 0.4655 - val_acc: 0.9031 945 Epoch 2972/5000 946 12s 152ms/step - loss: 0.1221 - acc: 0.9871 - val_loss: 0.4414 - val_acc: 0.9078 947 Epoch 2973/5000 948 12s 151ms/step - loss: 0.1232 - acc: 0.9867 - val_loss: 0.4539 - val_acc: 0.9062 949 Epoch 2974/5000 950 12s 152ms/step - loss: 0.1253 - acc: 0.9860 - val_loss: 0.4650 - val_acc: 0.9040 951 Epoch 2975/5000 952 12s 151ms/step - loss: 0.1249 - acc: 0.9865 - val_loss: 0.4454 - val_acc: 0.9089 953 Epoch 2976/5000 954 12s 151ms/step - loss: 0.1219 - acc: 0.9878 - val_loss: 0.4514 - val_acc: 0.9044 955 Epoch 2977/5000 956 12s 152ms/step - loss: 0.1255 - acc: 0.9853 - val_loss: 0.4585 - val_acc: 0.9056 957 Epoch 2978/5000 958 12s 151ms/step - loss: 0.1242 - acc: 0.9862 - val_loss: 0.4536 - val_acc: 0.9058 959 Epoch 2979/5000 960 12s 152ms/step - loss: 0.1248 - acc: 0.9864 - val_loss: 0.4638 - val_acc: 0.9048 961 Epoch 2980/5000 962 12s 150ms/step - loss: 0.1275 - acc: 0.9854 - val_loss: 0.4498 - val_acc: 0.9050 963 Epoch 2981/5000 964 12s 152ms/step - loss: 0.1260 - acc: 0.9856 - val_loss: 0.4647 - val_acc: 0.9033 965 Epoch 2982/5000 966 12s 151ms/step - loss: 0.1245 - acc: 0.9865 - val_loss: 0.4649 - val_acc: 0.9033 967 Epoch 2983/5000 968 12s 151ms/step - loss: 0.1265 - acc: 0.9857 - val_loss: 0.4429 - val_acc: 0.9103 969 Epoch 2984/5000 970 12s 151ms/step - loss: 0.1242 - acc: 0.9866 - val_loss: 0.4585 - val_acc: 0.9023 971 Epoch 2985/5000 972 12s 152ms/step - loss: 0.1257 - acc: 0.9858 - val_loss: 0.4490 - val_acc: 0.9038 973 Epoch 2986/5000 974 12s 151ms/step - loss: 0.1252 - acc: 0.9863 - val_loss: 0.4498 - val_acc: 0.9043 975 Epoch 2987/5000 976 12s 151ms/step - loss: 0.1248 - acc: 0.9864 - val_loss: 0.4459 - val_acc: 0.9058 977 Epoch 2988/5000 978 12s 152ms/step - loss: 0.1226 - acc: 0.9868 - val_loss: 0.4559 - val_acc: 0.9032 979 Epoch 2989/5000 980 12s 151ms/step - loss: 0.1237 - acc: 0.9868 - val_loss: 0.4560 - val_acc: 0.9051 981 Epoch 2990/5000 982 12s 151ms/step - loss: 0.1242 - acc: 0.9870 - val_loss: 0.4648 - val_acc: 0.9033 983 Epoch 2991/5000 984 12s 152ms/step - loss: 0.1245 - acc: 0.9863 - val_loss: 0.4604 - val_acc: 0.9018 985 Epoch 2992/5000 986 12s 151ms/step - loss: 0.1240 - acc: 0.9865 - val_loss: 0.4575 - val_acc: 0.9036 987 Epoch 2993/5000 988 13s 157ms/step - loss: 0.1226 - acc: 0.9872 - val_loss: 0.4626 - val_acc: 0.9028 989 Epoch 2994/5000 990 12s 152ms/step - loss: 0.1255 - acc: 0.9860 - val_loss: 0.4777 - val_acc: 0.8998 991 Epoch 2995/5000 992 12s 151ms/step - loss: 0.1256 - acc: 0.9860 - val_loss: 0.4574 - val_acc: 0.9023 993 Epoch 2996/5000 994 12s 151ms/step - loss: 0.1232 - acc: 0.9866 - val_loss: 0.4663 - val_acc: 0.9024 995 Epoch 2997/5000 996 12s 151ms/step - loss: 0.1210 - acc: 0.9881 - val_loss: 0.4663 - val_acc: 0.9046 997 Epoch 2998/5000 998 12s 152ms/step - loss: 0.1201 - acc: 0.9876 - val_loss: 0.4492 - val_acc: 0.9065 999 Epoch 2999/5000 1000 12s 152ms/step - loss: 0.1260 - acc: 0.9861 - val_loss: 0.4677 - val_acc: 0.9031 1001 Epoch 3000/5000 1002 12s 151ms/step - loss: 0.1256 - acc: 0.9861 - val_loss: 0.4517 - val_acc: 0.9044 1003 Epoch 3001/5000 1004 lr changed to 0.0009999999776482583 1005 12s 151ms/step - loss: 0.1226 - acc: 0.9877 - val_loss: 0.4332 - val_acc: 0.9071 1006 Epoch 3002/5000 1007 12s 151ms/step - loss: 0.1123 - acc: 0.9911 - val_loss: 0.4282 - val_acc: 0.9088 1008 Epoch 3003/5000 1009 12s 152ms/step - loss: 0.1072 - acc: 0.9926 - val_loss: 0.4277 - val_acc: 0.9110 1010 Epoch 3004/5000 1011 12s 152ms/step - loss: 0.1051 - acc: 0.9938 - val_loss: 0.4253 - val_acc: 0.9108 1012 Epoch 3005/5000 1013 12s 152ms/step - loss: 0.1041 - acc: 0.9941 - val_loss: 0.4242 - val_acc: 0.9101 1014 Epoch 3006/5000 1015 12s 151ms/step - loss: 0.1021 - acc: 0.9945 - val_loss: 0.4259 - val_acc: 0.9098 1016 Epoch 3007/5000 1017 12s 151ms/step - loss: 0.1034 - acc: 0.9940 - val_loss: 0.4255 - val_acc: 0.9100 1018 Epoch 3008/5000 1019 12s 152ms/step - loss: 0.1018 - acc: 0.9949 - val_loss: 0.4252 - val_acc: 0.9100 1020 Epoch 3009/5000 1021 12s 152ms/step - loss: 0.1029 - acc: 0.9945 - val_loss: 0.4276 - val_acc: 0.9103 1022 Epoch 3010/5000 1023 12s 151ms/step - loss: 0.1018 - acc: 0.9947 - val_loss: 0.4275 - val_acc: 0.9102 1024 Epoch 3011/5000 1025 12s 152ms/step - loss: 0.1004 - acc: 0.9951 - val_loss: 0.4237 - val_acc: 0.9106 1026 Epoch 3012/5000 1027 12s 152ms/step - loss: 0.0996 - acc: 0.9954 - val_loss: 0.4213 - val_acc: 0.9120 1028 Epoch 3013/5000 1029 12s 151ms/step - loss: 0.0997 - acc: 0.9953 - val_loss: 0.4247 - val_acc: 0.9112 1030 Epoch 3014/5000 1031 12s 151ms/step - loss: 0.0998 - acc: 0.9956 - val_loss: 0.4249 - val_acc: 0.9111 1032 Epoch 3015/5000 1033 12s 152ms/step - loss: 0.0999 - acc: 0.9953 - val_loss: 0.4261 - val_acc: 0.9103 1034 Epoch 3016/5000 1035 12s 151ms/step - loss: 0.0984 - acc: 0.9958 - val_loss: 0.4285 - val_acc: 0.9102 1036 Epoch 3017/5000 1037 12s 151ms/step - loss: 0.0999 - acc: 0.9954 - val_loss: 0.4284 - val_acc: 0.9098 1038 Epoch 3018/5000 1039 12s 154ms/step - loss: 0.0997 - acc: 0.9952 - val_loss: 0.4290 - val_acc: 0.9105 1040 Epoch 3019/5000 1041 12s 152ms/step - loss: 0.0992 - acc: 0.9955 - val_loss: 0.4273 - val_acc: 0.9118 1042 Epoch 3020/5000 1043 12s 151ms/step - loss: 0.0988 - acc: 0.9953 - val_loss: 0.4270 - val_acc: 0.9110 1044 Epoch 3021/5000 1045 12s 152ms/step - loss: 0.0988 - acc: 0.9957 - val_loss: 0.4298 - val_acc: 0.9104 1046 Epoch 3022/5000 1047 12s 151ms/step - loss: 0.0984 - acc: 0.9957 - val_loss: 0.4317 - val_acc: 0.9103 1048 Epoch 3023/5000 1049 12s 151ms/step - loss: 0.0976 - acc: 0.9960 - val_loss: 0.4282 - val_acc: 0.9107 1050 Epoch 3024/5000 1051 12s 152ms/step - loss: 0.0984 - acc: 0.9959 - val_loss: 0.4283 - val_acc: 0.9111 1052 Epoch 3025/5000 1053 12s 152ms/step - loss: 0.0969 - acc: 0.9960 - val_loss: 0.4288 - val_acc: 0.9090
过拟合依然严重，还是得继续减小网络。
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