本文共 27085 字，大约阅读时间需要 90 分钟。
Keras学习中文版
发现一个超详细代码介绍
原理部分见另外
import randomimport osos.environ['KERAS_BACKEND'] = 'tensorflow'from keras import backend as Kfrom keras.layers import Convolution2D, MaxPooling2Dfrom keras.layers import Dense, Dropout, Activation, Flattenfrom keras.models import Sequentialfrom keras.models import load_modelfrom keras.optimizers import SGDfrom keras.preprocessing.image import ImageDataGeneratorfrom keras.utils import np_utilsfrom sklearn.model_selection import train_test_splitfrom load_data import load_dataset, resize_image, IMAGE_SIZEclass Dataset: def __init__(self, path_name): # 训练集 self.train_images = None self.train_labels = None # 验证集 self.valid_images = None self.valid_labels = None # 测试集 self.test_images = None self.test_labels = None # 数据集加载路径 self.path_name = path_name # 当前库采用的维度顺序 self.input_shape = None self.nb_classes = None # 加载数据集并按照交叉验证的原则划分数据集并进行相关预处理工作 def load(self, img_rows=IMAGE_SIZE, img_cols=IMAGE_SIZE, img_channels=3): # 加载数据集到内存 images, labels, face_num = load_dataset(self.path_name) self.nb_classes = face_num train_images, valid_images, train_labels, valid_labels = train_test_split(images, labels, test_size=0.3, random_state=random.randint(0, 100)) _, test_images, _, test_labels = train_test_split(images, labels, test_size=0.5, random_state=random.randint(0, 100)) # 当前的维度顺序如果为'th'，则输入图片数据时的顺序为：channels,rows,cols，否则:rows,cols,channels # 这部分代码就是根据keras库要求的维度顺序重组训练数据集 #if K.image_dim_ordering() == 'th':#版本原因可能出错 if K.image_data_format() == 'channels_first': train_images = train_images.reshape(train_images.shape[0], img_channels, img_rows, img_cols) valid_images = valid_images.reshape(valid_images.shape[0], img_channels, img_rows, img_cols) test_images = test_images.reshape(test_images.shape[0], img_channels, img_rows, img_cols) self.input_shape = (img_channels, img_rows, img_cols) else: train_images = train_images.reshape(train_images.shape[0], img_rows, img_cols, img_channels) valid_images = valid_images.reshape(valid_images.shape[0], img_rows, img_cols, img_channels) test_images = test_images.reshape(test_images.shape[0], img_rows, img_cols, img_channels) self.input_shape = (img_rows, img_cols, img_channels) # 输出训练集、验证集、测试集的数量 print(train_images.shape[0], 'train samples') print(valid_images.shape[0], 'valid samples') print(test_images.shape[0], 'test samples') ''' 我们的模型使用categorical_crossentropy作为损失函数，因此需要根据类别数量nb_classes将 类别标签进行one-hot编码使其向量化，在这里我们的类别只有两种，经过转化后标签数据变为二维 ''' train_labels = np_utils.to_categorical(train_labels, self.nb_classes) valid_labels = np_utils.to_categorical(valid_labels, self.nb_classes) test_labels = np_utils.to_categorical(test_labels, self.nb_classes) # 像素数据浮点化以便归一化 train_images = train_images.astype('float32') valid_images = valid_images.astype('float32') test_images = test_images.astype('float32') # 将其归一化,图像的各像素值归一化到0~1区间 train_images /= 255 valid_images /= 255 test_images /= 255 self.train_images = train_images self.valid_images = valid_images self.test_images = test_images self.train_labels = train_labels self.valid_labels = valid_labels self.test_labels = test_labels# CNN网络模型类class Model: def __init__(self): self.model = None # 建立模型 def build_model(self, dataset, nb_classes=5): # 构建一个空的网络模型，它是一个线性堆叠模型，各神经网络层会被顺序添加，专业名称为序贯模型或线性堆叠模型 self.model = Sequential() # 以下代码将顺序添加CNN网络需要的各层，一个add就是一个网络层 self.model.add(Convolution2D(32, 3, 3, border_mode='same', input_shape=dataset.input_shape)) # 1 2维卷积层 卷积核 个数，大小，步长 self.model.add(Activation('relu')) # 2 激活函数层 self.model.add(Convolution2D(32, 3, 3)) # 3 2维卷积层 self.model.add(Activation('relu')) # 4 激活函数层 self.model.add(MaxPooling2D(pool_size=(2, 2))) # 5 池化层 self.model.add(Dropout(0.25)) # 6 Dropout层 self.model.add(Convolution2D(64, 3, 3, border_mode='same')) # 7 2维卷积层 self.model.add(Activation('relu')) # 8 激活函数层 self.model.add(Convolution2D(64, 3, 3)) # 9 2维卷积层 self.model.add(Activation('relu')) # 10 激活函数层 self.model.add(MaxPooling2D(pool_size=(2, 2))) # 11 池化层 self.model.add(Dropout(0.25)) # 12 Dropout层 self.model.add(Flatten()) # 13 Flatten层 self.model.add(Dense(512)) # 14 Dense层,又被称作全连接层 self.model.add(Activation('relu')) # 15 激活函数层 self.model.add(Dropout(0.5)) # 16 Dropout层 self.model.add(Dense(nb_classes)) # 17 Dense层 self.model.add(Activation('softmax')) # 18 分类层，输出最终结果 # 输出模型概况 self.model.summary() # 训练模型 def train(self, dataset, batch_size=20, nb_epoch=10, data_augmentation=True): sgd = SGD(lr=0.0007, decay=1e-6, momentum=0.9, nesterov=True) # 采用SGD+momentum的优化器进行训练，首先生成一个优化器对象 self.model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy']) # 完成实际的模型配置工作 # 不使用数据提升，所谓的提升就是从我们提供的训练数据中利用旋转、翻转、加噪声等方法创造新的 # 训练数据，有意识的提升训练数据规模，增加模型训练量 if not data_augmentation: self.model.fit(dataset.train_images, dataset.train_labels, batch_size=batch_size, nb_epoch=nb_epoch, validation_data=(dataset.valid_images, dataset.valid_labels), shuffle=True) # 使用实时数据提升 else: # 定义数据生成器用于数据提升，其返回一个生成器对象datagen，datagen每被调用一 # 次其生成一组数据（顺序生成），节省内存，其实就是python的数据生成器 datagen = ImageDataGenerator( featurewise_center=False, # 是否使输入数据去中心化（均值为0）， samplewise_center=False, # 是否使输入数据的每个样本均值为0 featurewise_std_normalization=False, # 是否数据标准化（输入数据除以数据集的标准差） samplewise_std_normalization=False, # 是否将每个样本数据除以自身的标准差 zca_whitening=False, # 是否对输入数据施以ZCA白化 rotation_range=20, # 数据提升时图片随机转动的角度(范围为0～180) width_shift_range=0.2, # 数据提升时图片水平偏移的幅度（单位为图片宽度的占比，0~1之间的浮点数） height_shift_range=0.2, # 同上，只不过这里是垂直 horizontal_flip=True, # 是否进行随机水平翻转 vertical_flip=False) # 是否进行随机垂直翻转 # 计算整个训练样本集的数量以用于特征值归一化、ZCA白化等处理 datagen.fit(dataset.train_images) # 利用生成器开始训练模型 self.model.fit_generator(datagen.flow(dataset.train_images, dataset.train_labels, batch_size=batch_size), samples_per_epoch=dataset.train_images.shape[0], nb_epoch=nb_epoch, validation_data=(dataset.valid_images, dataset.valid_labels)) MODEL_PATH = './model/face1.model' def save_model(self, file_path=MODEL_PATH): self.model.save(file_path) def load_model(self, file_path=MODEL_PATH): self.model = load_model(file_path) def evaluate(self, dataset): score = self.model.evaluate(dataset.test_images, dataset.test_labels, verbose=1) print("%s: %.2f%%" % (self.model.metrics_names[1], score[1] * 100)) # 识别人脸 def face_predict(self, image): # 依然是根据后端系统确定维度顺序 if K.image_dim_ordering() == 'th' and image.shape != (1, 3, IMAGE_SIZE, IMAGE_SIZE): image = resize_image(image) # 尺寸必须与训练集一致都应该是IMAGE_SIZE x IMAGE_SIZE image = image.reshape((1, 3, IMAGE_SIZE, IMAGE_SIZE)) # 与模型训练不同，这次只是针对1张图片进行预测 elif K.image_dim_ordering() == 'tf' and image.shape != (1, IMAGE_SIZE, IMAGE_SIZE, 3): image = resize_image(image) image = image.reshape((1, IMAGE_SIZE, IMAGE_SIZE, 3)) # 浮点并归一化 image = image.astype('float32') image /= 255 # 给出输入属于各个类别的概率 result_probability = self.model.predict_proba(image) print('result:', result_probability, max(result_probability[0])) # 给出类别预测：0-9 result = self.model.predict_classes(image) # 返回类别预测结果 return max(result_probability[0]),result[0]if __name__ == '__main__': dataset = Dataset('./data/') dataset.load() model = Model() model.build_model(dataset, dataset.nb_classes) model.train(dataset) model.save_model(file_path='./model/face1.model') model.evaluate(dataset) 训练 结果
374 train samples 161 valid samples 268 test samples Model: "sequential_1" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= conv2d_1 (Conv2D) (None, 64, 64, 32) 896 _________________________________________________________________ activation_1 (Activation) (None, 64, 64, 32) 0 _________________________________________________________________ conv2d_2 (Conv2D) (None, 62, 62, 32) 9248 _________________________________________________________________ activation_2 (Activation) (None, 62, 62, 32) 0 _________________________________________________________________ max_pooling2d_1 (MaxPooling2 (None, 31, 31, 32) 0 _________________________________________________________________ dropout_1 (Dropout) (None, 31, 31, 32) 0 _________________________________________________________________ conv2d_3 (Conv2D) (None, 31, 31, 64) 18496 _________________________________________________________________ activation_3 (Activation) (None, 31, 31, 64) 0 _________________________________________________________________ conv2d_4 (Conv2D) (None, 29, 29, 64) 36928 _________________________________________________________________ activation_4 (Activation) (None, 29, 29, 64) 0 _________________________________________________________________ max_pooling2d_2 (MaxPooling2 (None, 14, 14, 64) 0 _________________________________________________________________ dropout_2 (Dropout) (None, 14, 14, 64) 0 _________________________________________________________________ flatten_1 (Flatten) (None, 12544) 0 _________________________________________________________________ dense_1 (Dense) (None, 512) 6423040 _________________________________________________________________ activation_5 (Activation) (None, 512) 0 _________________________________________________________________ dropout_3 (Dropout) (None, 512) 0 _________________________________________________________________ dense_2 (Dense) (None, 2) 1026 _________________________________________________________________ activation_6 (Activation) (None, 2) 0 ================================================================= Total params: 6,489,634 Trainable params: 6,489,634 Non-trainable params: 0 Epoch 1/101/18 [>.............................] - ETA: 48s - loss: 0.6389 - accuracy: 0.6500
2/18 [==>...........................] - ETA: 27s - loss: 0.6374 - accuracy: 0.6250 3/18 [====>.........................] - ETA: 19s - loss: 0.6457 - accuracy: 0.6167 4/18 [=====>........................] - ETA: 14s - loss: 0.6392 - accuracy: 0.6250 5/18 [=======>......................] - ETA: 11s - loss: 0.6323 - accuracy: 0.6300 6/18 [=========>....................] - ETA: 9s - loss: 0.6434 - accuracy: 0.6083 7/18 [==========>...................] - ETA: 8s - loss: 0.6552 - accuracy: 0.6143 8/18 [============>.................] - ETA: 6s - loss: 0.6521 - accuracy: 0.6169 9/18 [==============>...............] - ETA: 5s - loss: 0.6560 - accuracy: 0.6149 10/18 [===============>..............] - ETA: 4s - loss: 0.6520 - accuracy: 0.6237 11/18 [=================>............] - ETA: 3s - loss: 0.6492 - accuracy: 0.6308 12/18 [===================>..........] - ETA: 3s - loss: 0.6476 - accuracy: 0.6282 13/18 [====================>.........] - ETA: 2s - loss: 0.6429 - accuracy: 0.6339 14/18 [======================>.......] - ETA: 2s - loss: 0.6354 - accuracy: 0.6423 15/18 [========================>.....] - ETA: 1s - loss: 0.6330 - accuracy: 0.6463 16/18 [=========================>....] - ETA: 0s - loss: 0.6348 - accuracy: 0.6401 17/18 [===========================>..] - ETA: 0s - loss: 0.6403 - accuracy: 0.6347 18/18 [==============================] - 9s 498ms/step - loss: 0.6350 - accuracy: 0.6412 - val_loss: 0.6009 - val_accuracy: 0.6584 Epoch 2/101/18 [>.............................] - ETA: 5s - loss: 0.6829 - accuracy: 0.6000
2/18 [==>...........................] - ETA: 5s - loss: 0.6944 - accuracy: 0.5500 3/18 [====>.........................] - ETA: 4s - loss: 0.6747 - accuracy: 0.5833 4/18 [=====>........................] - ETA: 4s - loss: 0.6857 - accuracy: 0.5625 5/18 [=======>......................] - ETA: 4s - loss: 0.6564 - accuracy: 0.6100 6/18 [=========>....................] - ETA: 3s - loss: 0.6321 - accuracy: 0.6417 7/18 [==========>...................] - ETA: 3s - loss: 0.6142 - accuracy: 0.6643 8/18 [============>.................] - ETA: 3s - loss: 0.6143 - accuracy: 0.6623 9/18 [==============>...............] - ETA: 2s - loss: 0.6040 - accuracy: 0.6724 10/18 [===============>..............] - ETA: 2s - loss: 0.6009 - accuracy: 0.6753 11/18 [=================>............] - ETA: 2s - loss: 0.6111 - accuracy: 0.6542 12/18 [===================>..........] - ETA: 1s - loss: 0.6159 - accuracy: 0.6453 13/18 [====================>.........] - ETA: 1s - loss: 0.6183 - accuracy: 0.6417 14/18 [======================>.......] - ETA: 1s - loss: 0.6116 - accuracy: 0.6496 15/18 [========================>.....] - ETA: 0s - loss: 0.6162 - accuracy: 0.6395 16/18 [=========================>....] - ETA: 0s - loss: 0.6103 - accuracy: 0.6433 17/18 [===========================>..] - ETA: 0s - loss: 0.6043 - accuracy: 0.6527 18/18 [==============================] - 6s 335ms/step - loss: 0.6079 - accuracy: 0.6469 - val_loss: 0.5700 - val_accuracy: 0.6584 Epoch 3/101/18 [>.............................] - ETA: 5s - loss: 0.7328 - accuracy: 0.5000
2/18 [==>...........................] - ETA: 5s - loss: 0.5827 - accuracy: 0.7250 3/18 [====>.........................] - ETA: 4s - loss: 0.5838 - accuracy: 0.7222 4/18 [=====>........................] - ETA: 4s - loss: 0.5677 - accuracy: 0.7162 5/18 [=======>......................] - ETA: 4s - loss: 0.5515 - accuracy: 0.7234 6/18 [=========>....................] - ETA: 3s - loss: 0.5562 - accuracy: 0.7018 7/18 [==========>...................] - ETA: 3s - loss: 0.5656 - accuracy: 0.6791 8/18 [============>.................] - ETA: 3s - loss: 0.5731 - accuracy: 0.6688 9/18 [==============>...............] - ETA: 2s - loss: 0.5767 - accuracy: 0.6552 10/18 [===============>..............] - ETA: 2s - loss: 0.5662 - accuracy: 0.6649 11/18 [=================>............] - ETA: 2s - loss: 0.5722 - accuracy: 0.6542 12/18 [===================>..........] - ETA: 1s - loss: 0.5742 - accuracy: 0.6538 13/18 [====================>.........] - ETA: 1s - loss: 0.5725 - accuracy: 0.6575 14/18 [======================>.......] - ETA: 1s - loss: 0.5704 - accuracy: 0.6642 15/18 [========================>.....] - ETA: 1s - loss: 0.5704 - accuracy: 0.6633 16/18 [=========================>....] - ETA: 0s - loss: 0.5646 - accuracy: 0.6688 17/18 [===========================>..] - ETA: 0s - loss: 0.5635 - accuracy: 0.6677 18/18 [==============================] - 6s 361ms/step - loss: 0.5663 - accuracy: 0.6667 - val_loss: 0.5245 - val_accuracy: 0.6584 Epoch 4/101/18 [>.............................] - ETA: 6s - loss: 0.5439 - accuracy: 0.6500
2/18 [==>...........................] - ETA: 7s - loss: 0.5967 - accuracy: 0.6000 3/18 [====>.........................] - ETA: 7s - loss: 0.5630 - accuracy: 0.6500 4/18 [=====>........................] - ETA: 6s - loss: 0.5582 - accuracy: 0.6500 5/18 [=======>......................] - ETA: 5s - loss: 0.5443 - accuracy: 0.6600 6/18 [=========>....................] - ETA: 5s - loss: 0.5199 - accuracy: 0.6917 7/18 [==========>...................] - ETA: 4s - loss: 0.5203 - accuracy: 0.6786 8/18 [============>.................] - ETA: 4s - loss: 0.5448 - accuracy: 0.6494 9/18 [==============>...............] - ETA: 3s - loss: 0.5587 - accuracy: 0.6322 10/18 [===============>..............] - ETA: 3s - loss: 0.5526 - accuracy: 0.6495 11/18 [=================>............] - ETA: 2s - loss: 0.5463 - accuracy: 0.6542 12/18 [===================>..........] - ETA: 2s - loss: 0.5350 - accuracy: 0.6709 13/18 [====================>.........] - ETA: 1s - loss: 0.5321 - accuracy: 0.6732 14/18 [======================>.......] - ETA: 1s - loss: 0.5393 - accuracy: 0.6642 15/18 [========================>.....] - ETA: 1s - loss: 0.5336 - accuracy: 0.6701 16/18 [=========================>....] - ETA: 0s - loss: 0.5266 - accuracy: 0.6752 17/18 [===========================>..] - ETA: 0s - loss: 0.5275 - accuracy: 0.6766 18/18 [==============================] - 7s 392ms/step - loss: 0.5204 - accuracy: 0.6864 - val_loss: 0.4376 - val_accuracy: 0.7516 Epoch 5/101/18 [>.............................] - ETA: 5s - loss: 0.4689 - accuracy: 0.7000
2/18 [==>...........................] - ETA: 5s - loss: 0.5072 - accuracy: 0.6500 3/18 [====>.........................] - ETA: 4s - loss: 0.4896 - accuracy: 0.7000 4/18 [=====>........................] - ETA: 4s - loss: 0.4796 - accuracy: 0.7125 5/18 [=======>......................] - ETA: 4s - loss: 0.4606 - accuracy: 0.7400 6/18 [=========>....................] - ETA: 4s - loss: 0.4284 - accuracy: 0.7750 7/18 [==========>...................] - ETA: 3s - loss: 0.4143 - accuracy: 0.7857 8/18 [============>.................] - ETA: 3s - loss: 0.4094 - accuracy: 0.7750 9/18 [==============>...............] - ETA: 3s - loss: 0.4267 - accuracy: 0.7500 10/18 [===============>..............] - ETA: 2s - loss: 0.4284 - accuracy: 0.7450 11/18 [=================>............] - ETA: 2s - loss: 0.4240 - accuracy: 0.7591 12/18 [===================>..........] - ETA: 2s - loss: 0.4187 - accuracy: 0.7792 13/18 [====================>.........] - ETA: 1s - loss: 0.4275 - accuracy: 0.7692 14/18 [======================>.......] - ETA: 1s - loss: 0.4220 - accuracy: 0.7857 15/18 [========================>.....] - ETA: 1s - loss: 0.4244 - accuracy: 0.7933 16/18 [=========================>....] - ETA: 0s - loss: 0.4263 - accuracy: 0.7969 17/18 [===========================>..] - ETA: 0s - loss: 0.4223 - accuracy: 0.8000 18/18 [==============================] - 7s 379ms/step - loss: 0.4186 - accuracy: 0.8083 - val_loss: 0.3016 - val_accuracy: 0.9876 Epoch 6/101/18 [>.............................] - ETA: 5s - loss: 0.3389 - accuracy: 1.0000
2/18 [==>...........................] - ETA: 5s - loss: 0.3579 - accuracy: 0.9250 3/18 [====>.........................] - ETA: 5s - loss: 0.3523 - accuracy: 0.9333 4/18 [=====>........................] - ETA: 4s - loss: 0.3374 - accuracy: 0.9375 5/18 [=======>......................] - ETA: 4s - loss: 0.3250 - accuracy: 0.9468 6/18 [=========>....................] - ETA: 4s - loss: 0.3273 - accuracy: 0.9298 7/18 [==========>...................] - ETA: 3s - loss: 0.3332 - accuracy: 0.9104 8/18 [============>.................] - ETA: 3s - loss: 0.3204 - accuracy: 0.9156 9/18 [==============>...............] - ETA: 3s - loss: 0.3251 - accuracy: 0.9023 10/18 [===============>..............] - ETA: 2s - loss: 0.3152 - accuracy: 0.9124 11/18 [=================>............] - ETA: 2s - loss: 0.3149 - accuracy: 0.9206 12/18 [===================>..........] - ETA: 2s - loss: 0.3111 - accuracy: 0.9188 13/18 [====================>.........] - ETA: 1s - loss: 0.3046 - accuracy: 0.9213 14/18 [======================>.......] - ETA: 1s - loss: 0.2935 - accuracy: 0.9270 15/18 [========================>.....] - ETA: 1s - loss: 0.2881 - accuracy: 0.9286 16/18 [=========================>....] - ETA: 0s - loss: 0.2808 - accuracy: 0.9299 17/18 [===========================>..] - ETA: 0s - loss: 0.2738 - accuracy: 0.9341 18/18 [==============================] - 7s 376ms/step - loss: 0.2713 - accuracy: 0.9322 - val_loss: 0.1445 - val_accuracy: 1.0000 Epoch 7/101/18 [>.............................] - ETA: 6s - loss: 0.2642 - accuracy: 0.9000
2/18 [==>...........................] - ETA: 5s - loss: 0.2516 - accuracy: 0.9000 3/18 [====>.........................] - ETA: 4s - loss: 0.2301 - accuracy: 0.9259 4/18 [=====>........................] - ETA: 4s - loss: 0.2276 - accuracy: 0.9324 5/18 [=======>......................] - ETA: 4s - loss: 0.2158 - accuracy: 0.9362 6/18 [=========>....................] - ETA: 3s - loss: 0.2158 - accuracy: 0.9386 7/18 [==========>...................] - ETA: 3s - loss: 0.2090 - accuracy: 0.9403 8/18 [============>.................] - ETA: 3s - loss: 0.2086 - accuracy: 0.9351 9/18 [==============>...............] - ETA: 2s - loss: 0.1988 - accuracy: 0.9425 10/18 [===============>..............] - ETA: 2s - loss: 0.1897 - accuracy: 0.9485 11/18 [=================>............] - ETA: 2s - loss: 0.1795 - accuracy: 0.9533 12/18 [===================>..........] - ETA: 1s - loss: 0.1819 - accuracy: 0.9530 13/18 [====================>.........] - ETA: 1s - loss: 0.1795 - accuracy: 0.9567 14/18 [======================>.......] - ETA: 1s - loss: 0.1791 - accuracy: 0.9562 15/18 [========================>.....] - ETA: 0s - loss: 0.1705 - accuracy: 0.9592 16/18 [=========================>....] - ETA: 0s - loss: 0.1642 - accuracy: 0.9618 17/18 [===========================>..] - ETA: 0s - loss: 0.1654 - accuracy: 0.9581 18/18 [==============================] - 6s 344ms/step - loss: 0.1640 - accuracy: 0.9598 - val_loss: 0.0638 - val_accuracy: 0.9938 Epoch 8/101/18 [>.............................] - ETA: 5s - loss: 0.0661 - accuracy: 1.0000
2/18 [==>...........................] - ETA: 4s - loss: 0.0745 - accuracy: 1.0000 3/18 [====>.........................] - ETA: 4s - loss: 0.0855 - accuracy: 1.0000 4/18 [=====>........................] - ETA: 4s - loss: 0.0884 - accuracy: 1.0000 5/18 [=======>......................] - ETA: 4s - loss: 0.0967 - accuracy: 0.9900 6/18 [=========>....................] - ETA: 3s - loss: 0.0848 - accuracy: 0.9917 7/18 [==========>...................] - ETA: 3s - loss: 0.0837 - accuracy: 0.9857 8/18 [============>.................] - ETA: 3s - loss: 0.0905 - accuracy: 0.9812 9/18 [==============>...............] - ETA: 2s - loss: 0.0837 - accuracy: 0.9828 10/18 [===============>..............] - ETA: 2s - loss: 0.0884 - accuracy: 0.9794 11/18 [=================>............] - ETA: 2s - loss: 0.0871 - accuracy: 0.9813 12/18 [===================>..........] - ETA: 2s - loss: 0.0876 - accuracy: 0.9829 13/18 [====================>.........] - ETA: 1s - loss: 0.0893 - accuracy: 0.9803 14/18 [======================>.......] - ETA: 1s - loss: 0.0862 - accuracy: 0.9818 15/18 [========================>.....] - ETA: 1s - loss: 0.0830 - accuracy: 0.9830 16/18 [=========================>....] - ETA: 0s - loss: 0.0841 - accuracy: 0.9809 17/18 [===========================>..] - ETA: 0s - loss: 0.0858 - accuracy: 0.9790 18/18 [==============================] - 7s 406ms/step - loss: 0.0831 - accuracy: 0.9802 - val_loss: 0.0338 - val_accuracy: 0.9938 Epoch 9/101/18 [>.............................] - ETA: 5s - loss: 0.0913 - accuracy: 0.9500
2/18 [==>...........................] - ETA: 5s - loss: 0.0608 - accuracy: 0.9750 3/18 [====>.........................] - ETA: 5s - loss: 0.0717 - accuracy: 0.9667 4/18 [=====>........................] - ETA: 4s - loss: 0.0645 - accuracy: 0.9750 5/18 [=======>......................] - ETA: 4s - loss: 0.0877 - accuracy: 0.9600 6/18 [=========>....................] - ETA: 4s - loss: 0.0920 - accuracy: 0.9583 7/18 [==========>...................] - ETA: 3s - loss: 0.0882 - accuracy: 0.9643 8/18 [============>.................] - ETA: 3s - loss: 0.0853 - accuracy: 0.9625 9/18 [==============>...............] - ETA: 3s - loss: 0.0846 - accuracy: 0.9667 10/18 [===============>..............] - ETA: 3s - loss: 0.0890 - accuracy: 0.9650 11/18 [=================>............] - ETA: 2s - loss: 0.1150 - accuracy: 0.9409 12/18 [===================>..........] - ETA: 2s - loss: 0.1293 - accuracy: 0.9375 13/18 [====================>.........] - ETA: 1s - loss: 0.1628 - accuracy: 0.9291 14/18 [======================>.......] - ETA: 1s - loss: 0.2432 - accuracy: 0.9015 15/18 [========================>.....] - ETA: 1s - loss: 0.2775 - accuracy: 0.8946 16/18 [=========================>....] - ETA: 0s - loss: 0.2833 - accuracy: 0.8917 17/18 [===========================>..] - ETA: 0s - loss: 0.2679 - accuracy: 0.8982 18/18 [==============================] - 8s 424ms/step - loss: 0.2613 - accuracy: 0.9011 - val_loss: 0.0490 - val_accuracy: 0.9752 Epoch 10/101/18 [>.............................] - ETA: 7s - loss: 0.0377 - accuracy: 1.0000
2/18 [==>...........................] - ETA: 6s - loss: 0.0480 - accuracy: 1.0000 3/18 [====>.........................] - ETA: 6s - loss: 0.0612 - accuracy: 1.0000 4/18 [=====>........................] - ETA: 5s - loss: 0.0713 - accuracy: 0.9875 5/18 [=======>......................] - ETA: 5s - loss: 0.0682 - accuracy: 0.9900 6/18 [=========>....................] - ETA: 5s - loss: 0.0671 - accuracy: 0.9917 7/18 [==========>...................] - ETA: 4s - loss: 0.0813 - accuracy: 0.9857 8/18 [============>.................] - ETA: 4s - loss: 0.0763 - accuracy: 0.9875 9/18 [==============>...............] - ETA: 3s - loss: 0.0720 - accuracy: 0.9889 10/18 [===============>..............] - ETA: 3s - loss: 0.0726 - accuracy: 0.9850 11/18 [=================>............] - ETA: 3s - loss: 0.0722 - accuracy: 0.9818 12/18 [===================>..........] - ETA: 2s - loss: 0.0713 - accuracy: 0.9833 13/18 [====================>.........] - ETA: 2s - loss: 0.0696 - accuracy: 0.9846 14/18 [======================>.......] - ETA: 1s - loss: 0.0727 - accuracy: 0.9821 15/18 [========================>.....] - ETA: 1s - loss: 0.0695 - accuracy: 0.9833 16/18 [=========================>....] - ETA: 0s - loss: 0.0669 - accuracy: 0.9844 17/18 [===========================>..] - ETA: 0s - loss: 0.0657 - accuracy: 0.9853 18/18 [==============================] - 8s 444ms/step - loss: 0.0632 - accuracy: 0.9861 - val_loss: 0.0322 - val_accuracy: 0.993832/268 [==>...........................] - ETA: 0s
64/268 [======>.......................] - ETA: 0s 96/268 [=========>....................] - ETA: 0s 128/268 [=============>................] - ETA: 0s 160/268 [================>.............] - ETA: 0s 192/268 [====================>.........] - ETA: 0s 224/268 [========================>.....] - ETA: 0s 256/268 [===========================>..] - ETA: 0s 268/268 [==============================] - 1s 3ms/step accuracy: 99.25%转载地址：http://fdupz.baihongyu.com/