微调模型删除之前添加的层

Question

我使用 Keras 2.2.4。我训练了一个模型，我希望每 30 个 epoch 使用新的数据内容（图像分类）对其进行微调。

每天我都会在类中添加更多图像来提供模型。每 30 个 epoch 对模型进行一次重新训练。 我使用 2 个条件，第一个条件是之前没有训练过模型，第二个条件是模型已经训练过，然后我想用新的内容/类对其进行微调。

model_base = keras.applications.vgg19.VGG19(include_top=False, input_shape=(*IMG_SIZE, 3), weights='imagenet')
output = GlobalAveragePooling2D()(model_base.output)

# If we resume a pretrained model load it
if os.path.isfile(os.path.join(MODEL_PATH, 'weights.h5')): 
    print('Using existing weights...')
    base_lr = 0.0001

    model = load_model(os.path.join(MODEL_PATH, 'weights.h5'))
    output = Dense(len(all_character_names), activation='softmax', name='d2')(output)
    model = Model(model_base.input, output)

    for layer in model_base.layers[:-2]:
        layer.trainable = False 
else:
    base_lr = 0.001

    output = BatchNormalization()(output)
    output = Dropout(0.5)(output)
    output = Dense(2048, activation='relu', name='d1')(output)
    output = BatchNormalization()(output)
    output = Dropout(0.5)(output)
    output = Dense(len(all_character_names), activation='softmax', name='d2')(output)
    model = Model(model_base.input, output)

    for layer in model_base.layers[:-5]:
        layer.trainable = False 


opt = optimizers.Adam(lr=base_lr, decay=base_lr / epochs)
model.compile(optimizer=opt,
            loss='categorical_crossentropy',
            metrics=['accuracy'])

第一次模型总结：

...
_________________________________________________________________
block5_conv4 (Conv2D)        (None, 14, 14, 512)       2359808   
_________________________________________________________________
block5_pool (MaxPooling2D)   (None, 7, 7, 512)         0         
_________________________________________________________________
global_average_pooling2d_1 ( (None, 512)               0         
_________________________________________________________________
batch_normalization_1 (Batch (None, 512)               2048      
_________________________________________________________________
dropout_1 (Dropout)          (None, 512)               0         
_________________________________________________________________
d1 (Dense)                   (None, 2048)              1050624   
_________________________________________________________________
batch_normalization_2 (Batch (None, 2048)              8192      
_________________________________________________________________
dropout_2 (Dropout)          (None, 2048)              0         
_________________________________________________________________
d2 (Dense)                   (None, 19)                38931     
=================================================================
Total params: 21,124,179
Trainable params: 10,533,907
Non-trainable params: 10,590,272

模型总结第二次：

...
_________________________________________________________________
block5_conv4 (Conv2D)        (None, 14, 14, 512)       2359808   
_________________________________________________________________
block5_pool (MaxPooling2D)   (None, 7, 7, 512)         0         
_________________________________________________________________
global_average_pooling2d_1 ( (None, 512)               0         
_________________________________________________________________
d2 (Dense)                   (None, 19)                9747      
=================================================================
Total params: 20,034,131
Trainable params: 2,369,555
Non-trainable params: 17,664,576

问题：当模型存在并加载以进行微调时，它似乎已松散了第一次添加的所有附加层（Dense 2048、Dropout 等）

我需要再次添加这些图层吗？这似乎没有任何意义，因为它会丢失第一次传递的训练信息。

注意：我可能不需要设置 base_lr，因为保存模型还应该保存之前停止状态的学习率，但我稍后会检查。

Answer 1

请注意，加载模型后：

model = load_model(os.path.join(MODEL_PATH, 'weights.h5'))

你不使用它。你只是再次覆盖它

model = Model(model_base.input, output)

其中输出也被定义为对 base_model 的操作。 在我看来，您只想删除 load_model 之后的行。