無法使用ResNet50為Keras中的微調加載重量

Question

我首先使用以下方法在我的數據集上凍結了ResNet-50圖層：

model_r50 = ResNet50(weights='imagenet', include_top=False)
model_r50.summary()

input_layer = Input(shape=(img_width,img_height,3),name = 'image_input')

output_r50 = model_r50(input_layer)

fl = Flatten(name='flatten')(output_r50)
dense = Dense(1024, activation='relu', name='fc1')(fl)
drop = Dropout(0.5, name='drop')(dense)
pred = Dense(nb_classes, activation='softmax', name='predictions')(drop)
fine_model = Model(outputs=pred,inputs=input_layer)
for layer in model_r50.layers:
    layer.trainable = False
    print layer

fine_model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
fine_model.summary()

然后，我嘗試使用以下方法對圖層解凍進行微調：

model_r50 = ResNet50(weights='imagenet', include_top=False)
model_r50.summary()

input_layer = Input(shape=(img_width,img_height,3),name = 'image_input')

output_r50 = model_r50(input_layer)

fl = Flatten(name='flatten')(output_r50)
dense = Dense(1024, activation='relu', name='fc1')(fl)
drop = Dropout(0.5, name='drop')(dense)
pred = Dense(nb_classes, activation='softmax', name='predictions')(drop)
fine_model = Model(outputs=pred,inputs=input_layer)
weights = 'val54_r50.01-0.86.hdf5'
fine_model.load_weights('models/'+weights)
fine_model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
fine_model.summary()

但是我無處可去。 我剛剛解凍網絡並沒有改變任何東西！

  load_weights_from_hdf5_group(f, self.layers)
  File "/usr/local/lib/python2.7/dist-packages/keras/engine/topology.py", line 3008, in load_weights_from_hdf5_group
    K.batch_set_value(weight_value_tuples)
  File "/usr/local/lib/python2.7/dist-packages/keras/backend/tensorflow_backend.py", line 2189, in batch_set_value
    get_session().run(assign_ops, feed_dict=feed_dict)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 778, in run
    run_metadata_ptr)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 961, in _run
    % (np_val.shape, subfeed_t.name, str(subfeed_t.get_shape())))
ValueError: Cannot feed value of shape (128,) for Tensor u'Placeholder_140:0', which has shape '(512,)'

而且它並不一致。 我大部分時間都有不同的形狀。 為什么會這樣？ 如果我只是將ResNet更改為VGG19，則不會發生這種情況。 Keras的ResNet有問題嗎？

Answer 1

你的fine_model是一個Model ，里面有另一個Model （即ResNet50 ）。 似乎問題是save_weight()和load_weight()無法正確處理這種類型的嵌套Model 。

也許您可以嘗試以不會導致“嵌套Model ”的方式構建模型。 例如，

input_layer = Input(shape=(img_width, img_height, 3), name='image_input')
model_r50 = ResNet50(weights='imagenet', include_top=False, input_tensor=input_layer)
output_r50 = model_r50.output
fl = Flatten(name='flatten')(output_r50)
...

Answer 2

以下程序通常對我有用：

1. 將權重加載到凍結模型中。

2. 將圖層更改為可訓練。

3. 編譯模型。

即在這種情況下：

model_r50 = ResNet50(weights='imagenet', include_top=False)
model_r50.summary()

input_layer = Input(shape=(img_width,img_height,3),name = 'image_input')

output_r50 = model_r50(input_layer)

fl = Flatten(name='flatten')(output_r50)
dense = Dense(1024, activation='relu', name='fc1')(fl)
drop = Dropout(0.5, name='drop')(dense)
pred = Dense(nb_classes, activation='softmax', name='predictions')(drop)
fine_model = Model(outputs=pred,inputs=input_layer)
for layer in model_r50.layers:
    layer.trainable = False
    print layer

weights = 'val54_r50.01-0.86.hdf5'
fine_model.load_weights('models/'+weights)

for layer in model_r50.layers:
    layer.trainable = True

fine_model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
fine_model.summary()