zoeygxy opened a new pull request #15959: Numpy compatible logaddexp2 URL: https://github.com/apache/incubator-mxnet/pull/15959 ## Description ## - Added [np.logaddexp2](https://docs.scipy.org/doc/numpy/reference/generated/numpy.logaddexp2.html) and its docs for future rendering. Scalar calculation is supported. - Fixed zero-size tensor error in BinaryBroadcastBackwardUseIn (GPU version). ## Checklist ## ### Essentials ### - [x] Changes are complete (i.e. I finished coding on this PR) - [x] All changes have test coverage: - Unit tests are added for small changes to verify correctness - [x] Code is well-documented: - For user-facing API changes, API doc string has been updated. - [x] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change ### Changes ### - [x] Added [np.logaddexp2](https://docs.scipy.org/doc/numpy/reference/generated/numpy.logaddexp2.html) and its docs for future rendering. Scalar calculation is supported. - [x] Fixed zero-size tensor error in BinaryBroadcastBackwardUseIn. ## Comments ## - The previous error case description: when one operand is non-zero-size (eg. shape of (1, 3)) and the other is zero-size (eg. shape of (2, 0, 3)), the backward calculation on GPU will raise a floating point error. Temporarily fixed in this PR as well as PR numpy compatible logaddexp.
---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services
