frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update torch.compile issue template (#150192)

Browse Source 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150192
Approved by: https://github.com/malfet
ghstack dependencies: #149947
This commit is contained in:
eellison
2025-03-28 08:44:12 -07:00
committed by PyTorch MergeBot
parent f04cf13bdd
commit 15dbad2115
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
.github/ISSUE_TEMPLATE/pt2-bug-report.yml vendored
View File

@ -20,7 +20,7 @@ body:
- Don't compare indices of max/min etc, because that avoids the above requirement
- If comparing eager and torch.compile at fp16/bf16, you should use fp32 as baseline
- When comparing eager and torch.compile, use a higher precision result as a baseline. `torch._dynamo.utils.same` with fp64_ref will handle this comparison.
- Ensure rng state used to compare results is equivalent. Use `torch._inductor.config.fallback_random=True` and reset the torch rng seed between comparisons