-
Notifications
You must be signed in to change notification settings - Fork 24.7k
fix sigmoid for torch.complex datatypes on CPU #140391
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/140391
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 7edb584 with merge base 6cb186e ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This just undoes the vectorization. Is it hard to fix the vectorized implementation? |
Hi @ezyang , I think this is not easy. The problem is that in the scalar implementation of asin there are special handling for the case where real/imag numbers are 0/Inf/NaN, but these special handling for 0/Inf/NaN are missing in the vectorized implementation, which is not easy to do. For correctness, I will temporarily revert the implementation of asin to a scalar implementation. |
hi @Skylion007 , could you please review this PR? thanks! |
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Stack from ghstack (oldest at bottom):
Fix #135777.
This issue is caused by the lack of special handling of the case where the real number/imag number is 0/Inf/NaN in the vectorized implementation of
reciprocal
. For correctness, I temporarily fallback the implementation ofreciprocal
to scalar implementation.cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @ezyang @anjali411 @dylanbespalko @mruberry @nikitaved @amjames