wbhu/DnCNN-tensorflow

DnCNN last layer

eastchun opened this issue · 1 comments

eastchun commented

DnCNN last layer: with tf.variable_scope('block17'): ---> ('block20')

Ant-dong commented

why not block21 in DnCNN last layer. I guess that author can increase layer in debug, but there is not change in the output layer.