oleg-trott edited a comment on issue #17684: The output of the ReLU layer in MXNET is different from that in tensorflow and cntk URL: https://github.com/apache/incubator-mxnet/issues/17684#issuecomment-592997242 @braindotai > As given [here](https://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html) make sure that you are normalizing your image as below I don't think the normalization is the culprit, if the previous layer outputs match. Keras probably uses the same network weights with all backends. @Justobe I don't have the other frameworks installed, so I can't reproduce this, but my suggestion is: check the inputs to `relu`. Since it's a very simple function ``` x * (x > 0) ``` it should be easy to check that the output is what it's supposed to be. If it is not, use the input and output of `conv1_relu` to try to create a reproducible case that doesn't need other frameworks.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
