ValueError: The following Variables were created within a Lambda layer | tensorflow2.3 | Mask-RCNN

原创于 2024-11-02 22:51:23 发布 · 466 阅读

4 ·

CC 4.0 BY-SA版权

文章标签：

#tensorflow #人工智能 #python

亲测：在Tensorflow2.3的环境下，网上的办法全部无用，经查官方文档后修改如下：

有Tensorflow2.3兼容性问题的Mask RCNN的github项目：

https://github.com/matterport/Mask_RCNN

报错信息：

ValueError: The following Variables were created within a Lambda layer (anchors) but are not tracked by said layer: <tf.Variable ‘anchors/Variable:0’ shape=(1, 261888, 4) dtype=float32> The layer cannot safely ensure proper Variable reuse across multiple calls, and consquently this behavior is disallowed for safety. Lambda layers are not well suited to stateful computation; instead, writing a subclassed Layer is the recommend way to define layers with Variables.

问题原因：

从Tensorflow2.3开始，由于安全性原因，官方不推荐使用Lambda layer，需要使用subclass（Layer子类）方式实现。

将：

anchors = KL.Lambda(lambda x: tf.Variable(anchors), name="anchors")(input_image)

改为：

class ConstLayer(tf.keras.layers.Layer):
                def __init__(self, x, name=None):
                    super(ConstLayer, self).__init__(name=name)
                    self.x = tf.Variable(x)

                def call(self, input):
                    return self.x

            anchors = ConstLayer(anchors, name="anchors")(input_image)