Question

我正在尝试从以下Dense图层获取权重：

x = Dense(1024)(Flatten()(previous_layer))

如果我尝试执行以下操作：

x = Dense(1024)
weights = x.get_weights()

这很好用，但我的理解是这些权重将无用，因为我们没有向图层提供任何输入。

但是，如果我尝试执行以下操作：

x = Dense(1024)(Flatten()(previous_layer))
weights = x.get_weights()

这不起作用，因为x现在是Tensor对象，并且没有get_weights方法：

'Tensor' object has no attribute 'get_weights'

我做错了什么？

Answer 1

将图层应用于某个输入张量（Dense(n)）时，图层（Dense(n)(input)）与输出张量之间存在差异。您需要将图层存储在变量中，而不仅仅是输出张量：

>>> import keras
>>> input_layer = keras.layers.Input((2,))
>>> layer = keras.layers.Dense(3) # create a layer
>>> print(layer)
<keras.layers.core.Dense object at 0x7f03ca9d4d68>
>>> print(layer.get_weights()) # the layer does not have weights yet
[]
>>> output_tensor = layer(input_layer) # apply the layer to the input tensor
>>> print(output_tensor)
Tensor("dense_1/BiasAdd:0", shape=(?, 3), dtype=float32)
>>> print(layer.get_weights()) # now get the weights
[array([[-0.84973848, -0.19682372, -0.14602524],
       [ 0.70318353, -0.1578933 , -0.94751853]], dtype=float32),
 array([ 0.,  0.,  0.], dtype=float32)]

Answer 2

lstm_output = keras.layers.LSTM(100,return_sequences = True)(x_5)

output_location_final = Dense(2,activation='tanh')

#lstm_output is a tensor

result = output_location_final(lstm_output)

#result is a layer

weight = output_location_final.get_weights()

print(weight)

从Keras图层获取权重

2 个答案: