rahul003 opened a new pull request #10435: [MXNET-289] Allow specifying number 
of batches to run in an epoch to fix hang in dist sync
URL: https://github.com/apache/incubator-mxnet/pull/10435
 
 
   ## Description ##
   https://issues.apache.org/jira/browse/MXNET-289
   
   Fixes https://github.com/apache/incubator-mxnet/issues/9611
   
   The problem is that different parts of the dataset being worked on different 
machines can have different number of batches when the number of examples is 
not perfectly divisible. Some machines could get more one more batch than some 
other machines, but there's no progress because in dist_sync mode we wait for 
all workers to send gradients.
   
   This fix is similar to the argument for the older model API.
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to 
the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) 
created (except PRs with tiny changes)
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   - [ ] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [ ] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, README.md is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - Check the API doc at 
http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
   - [ ] To the my best knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change
   
   ### Changes ###
   - [ ] Add epoch_size arg to Module.fit
   - [ ] Updated image-classification examples to use this argument
   - [ ] Updated division in those examples to support both py2 and py3 by 
casting the result of division to int (now a/b can be float)
   
   ## Comments ##
   - If this change is a backward incompatible change, why must this change be 
made.
   - Interesting edge cases to note here
   

----------------------------------------------------------------
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:
us...@infra.apache.org


With regards,
Apache Git Services

Reply via email to