我正在用14个新类对MobileNet进行微调。当我通过以下方式添加新图层时:
x=mobile.layers[-6].output
x=Flatten(x)
predictions = Dense(14, activation='softmax')(x)
model = Model(inputs=mobile.input, outputs=predictions)
我得到了错误:
'Tensor' object has no attribute 'lower'
也使用:
model.compile(Adam(lr=.0001), loss='categorical_crossentropy', metrics=['accuracy'])
model.fit_generator(train_batches, steps_per_epoch=18,
validation_data=valid_batches, validation_steps=3, epochs=60, verbose=2)
我得到了错误:
Error when checking target: expected dense_1 to have 4 dimensions, but got array with shape (10, 14)
lower
是什么意思?我看到了其他微调脚本,除了模型名称(在这种情况下为X)之外,没有其他参数。
答案 0 :(得分:8)
张量必须在调用时传递给图层,而不是作为参数传递。因此,它必须是这样的:
x = Flatten()(x) # first the layer is constructed and then it is called on x
更清楚地说,它等效于此:
flatten_layer = Flatten() # instantiate the layer
x = flatten_layer(x) # call it on the given tensor