amiithinks / amii-tf-nn Goto Github PK

It would be great to use the layers from tf.layers but they reset the name scope before their components are created, which doesn't really make sense to me. For example,

with tf.name_scope(name):
    preactivate = tf.layers.dense(
        inputs,
        units,
        # TODO This doesn't work, how should the scope be set?
        # scope=name
    )

will not put this layer under the name scope. Is this desired behaviour, a bug, or is there a way around this?

Should maybe ask this on StackOverflow.