Add more unit tests for both distributed and extensions

ROCm / apex

A PyTorch Extension: Tools for easy mixed precision and distributed training in Pytorch

BSD 3-Clause "New" or "Revised" License

19 stars 17 forks source link

Add more unit tests for both distributed and extensions #58

Closed hubertlu-tw closed 2 years ago

hubertlu-tw commented 2 years ago

Add more unit tests for extensions and distributed to ensure the correctness. CI requires to be refactored accordingly.

jithunnair-amd commented 2 years ago

@hubertlu-tw Can you please merge from master into this branch? It will remove changes from the diff which were already part of merged PR #57, giving a cleaner diff.

hubertlu-tw commented 2 years ago

jenkins: retest this please