deutschmn commented on pull request #925: URL: https://github.com/apache/systemml/pull/925#issuecomment-636802804
Hi @j143, thanks for having a look at my PR! Sorry, I realised I still had a very low number for the negative samples in the Jupyter notebook. That caused the network to learn that basically almost every label should be 1 and hence immediately reached nearly zero loss. I corrected the issue in the most recent commit and in the Colab notebook 😊 Not sure, if that's the same for everyone though. Cheers, Patrick ---------------------------------------------------------------- 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
