lmmcc commented on issue #11897: why fixed_param_names in module have no use URL: https://github.com/apache/incubator-mxnet/issues/11897#issuecomment-408292452 @wkcn yes,i use your method,and the parameter is fixed..but when i use the new trained model to extract feature using layer 'fc1' which is before fc7 layer，the result is different from the pre-trained one ,it is so strange....and i use modle-slim.py to get the same layer betweent these two models...the *.param files is diff......their parameters is same...then which part is different.......
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
