tensorflow - 与普通层不同，Keras 自定义层不返回权重

Question

我正在尝试获取层的权重。当使用 keras 层并将输入连接到它时，它似乎可以正常工作。 但是，在将它包装到我的自定义层中时，它不再起作用了。这是错误还是我遗漏了什么？

编辑:注意事项:

我读到可以在自定义层的 build() 中定义可训练变量。但是，由于自定义层由密集的 keras 层组成(之后可能还有更多的 keras 层)，因此这些层应该已经定义了可训练变量和权重/偏差初始化器。 (在 TestLayer 的 init() 中，我看不到用在 TestLayer 的 build() 中定义的变量覆盖它们的方法。

class TestLayer(layers.Layer):
    def __init__(self):
        super(TestLayer, self).__init__()
        self.test_nn = layers.Dense(3)

    def build(self, input_shape):
        super(TestLayer, self).build(input_shape)


    def call(self, inputs, **kwargs):
        test_out = test_nn(inputs) # which is test_in
        return test_out


test_in = layers.Input((2,))
test_nn = layers.Dense(3)
print(test_nn.get_weights()) # empty, since no connection to the layer
test_out = test_nn(test_in)
print(test_nn.get_weights()) # layer returns weights+biases

testLayer = TestLayer()
features = testLayer(test_in)
print(testLayer.get_weights()) # Problem: still empty, even though connected to input.

Answer 1

documentation说 build() 方法应该调用 add_weight() 而你没有:

Should have the calls to add_weight(), and then call the super's build()

如果您继承layers.Layer，您也不需要在您的类中定义密集层。 这是你应该如何子类化:

import tensorflow as tf
from tensorflow.keras import layers

class TestLayer(layers.Layer):
    def __init__(self, outshape=3):
        super(TestLayer, self).__init__()
        self.outshape = outshape

    def build(self, input_shape):
        self.kernel = self.add_weight(name='kernel',
                                      shape=(int(input_shape[1]), self.outshape),
                                      trainable=True)

        super(TestLayer, self).build(input_shape)


    def call(self, inputs, **kwargs):
        return tf.matmul(inputs, self.kernel)

test_in = layers.Input((2,))

testLayer = TestLayer()
features = testLayer(test_in)
print(testLayer.get_weights())
#[array([[-0.68516827, -0.01990592,  0.88364804],
#       [-0.459718  ,  0.19161093,  0.39982545]], dtype=float32)]

Here是子类化 Layer 类的更多示例。

但是，如果您坚持按照自己的方式实现它并且想使用get_weights()，则必须覆盖它(在这种情况下您可以创建一个没有子类的类):

import tensorflow as tf
from tensorflow.keras import layers

class TestLayer(layers.Layer):
    def __init__(self, outshape=3):
        super(TestLayer, self).__init__()
        self.test_nn = layers.Dense(outshape)
        self.outshape = outshape

    def build(self, input_shape):
        super(TestLayer, self).build(input_shape)

    def call(self, inputs, **kwargs):
        return self.test_nn(inputs)

    def get_weights(self):
        with tf.Session() as sess:
            sess.run([x.initializer for x in self.test_nn.trainable_variables])
            return sess.run(self.test_nn.trainable_variables)

test_in = layers.Input((2,))

testLayer = TestLayer()
features = testLayer(test_in)
print(testLayer.get_weights())
#[array([[ 0.5692867 ,  0.726858  ,  0.37790012],
#       [ 0.2897135 , -0.7677493 , -0.58776844]], dtype=float32), #array([0., 0., 0.], dtype=float32)]