MonicaGu removed 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-603136144 I ran your code and reproduced the difference in ReLU layer. However, when I use the mxnet output of `conv1_bn` as the input of ReLU, TensorFlow and mxnet backends have the same output. This is my code: ``` import pickle import numpy as np import tensorflow as tf import mxnet as mx with open("output/{}_{}.pkl".format("mxnet",imagename[:-4]), "rb+") as f: backend_output_dict = pickle.load(f) count = 0 for each in backend_output_dict: count += 1 if count == 3: input = each break mxoutput = mx.nd.relu(mx.nd.array(input)) tfoutput = tf.nn.relu(input) print(delta(mxoutput.asnumpy(), tfoutput)) ``` The output of running the code is: ``` [0.] [0.] ``` So I suppose there might be a problem in BatchNorm and the slight difference in `conv1_bn` might lead to a big difference in `conv1_relu`.
---------------------------------------------------------------- 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
