Description
PyTorch before 3.7.0 has a bernoulli_p decompose function in decompositions.py even though it lacks full consistency with the eager CPU implementation, negatively affecting nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d for fallback_random=True.
Metrics
CVSS 3.1 5.3/10
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:NWeakness enumeration
Affected software
| Vendor | Product | Versions |
|---|---|---|
| n/a | n/a | n/a |
References
- https://gist.github.com/shaoyuyoung/4bcefba4004f8271e64b5185c95a248aThird Party Advisory
- https://gist.github.com/shaoyuyoung/e636f2e7a306105b7e96809e2b85c28aThird Party Advisory
- https://github.com/pytorch/pytorch/compare/v2.6.0...v2.7.0Product
- https://github.com/pytorch/pytorch/issues/142853Issue Tracking
- https://github.com/pytorch/pytorch/pull/143460Issue Tracking, Patch
Timeline
- Published
- Sep 25, 2025
- Last modified
- Nov 5, 2026
- NVD status
- Analyzed