Keras Transfer-Learning设置层.可训练到True没有效果

Question

我想使用tf微调效率。Keras (tensorflow 2.3)，但我不能正确地改变层的训练状态。我的模型是这样的:

data_augmentation_layers = tf.keras.Sequential([
keras.layers.experimental.preprocessing.RandomFlip("horizontal_and_vertical"),
keras.layers.experimental.preprocessing.RandomRotation(0.8)])
efficientnet = EfficientNetB3(weights="imagenet", include_top=False,
input_shape=(*img_size, 3))
#Setting to not trainable as described in the standard keras FAQ
efficientnet.trainable = False
inputs = keras.layers.Input(shape=(*img_size, 3))
augmented = augmentation_layers(inputs)
base = efficientnet(augmented, training=False)
pooling = keras.layers.GlobalAveragePooling2D()(base)
outputs = keras.layers.Dense(5, activation="softmax")(pooling)
model = keras.Model(inputs=inputs, outputs=outputs)
model.compile(loss="categorical_crossentropy", optimizer=keras_opt, metrics=["categorical_accuracy"])

这样做是为了让我在自定义顶部的随机权重不会尽快破坏权重。

Model: "functional_1"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
input_2 (InputLayer)         [(None, 512, 512, 3)]     0         
_________________________________________________________________
sequential (Sequential)      (None, 512, 512, 3)       0         
_________________________________________________________________
efficientnetb3 (Functional)  (None, 16, 16, 1536)      10783535  
_________________________________________________________________
global_average_pooling2d (Gl (None, 1536)              0         
_________________________________________________________________
dense (Dense)                (None, 5)                 7685      
=================================================================
Total params: 10,791,220
Trainable params: 7,685
Non-trainable params: 10,783,535

在此之前一切似乎都很顺利。我训练我的模型2个时代，然后我想开始微调有效率的基础。因此我调用

for l in model.get_layer("efficientnetb3").layers:
if not isinstance(l, keras.layers.BatchNormalization):
l.trainable = True
model.compile(loss="categorical_crossentropy", optimizer=keras_opt, metrics=["categorical_accuracy"])

我重新编译并再次打印摘要，看到不可训练权重的数量保持不变。而且拟合效果也不如冷冻效果好。

dense (Dense)                (None, 5)                 7685      
=================================================================
Total params: 10,791,220
Trainable params: 7,685
Non-trainable params: 10,783,535

Ps:我也试过efficientnet3.trainable = True，但这也没有效果。

这可能与我同时使用顺序模型和功能模型的事实有关吗?

Answer 1

对我来说，问题是对模型的一部分使用顺序API。当我更改为顺序时，我的model. summary()显示所有子层，并且可以将其中一些设置为可训练的，而其他的则不可训练。