CVE-2025-46153
BaseFortify
Publication date: 2025-09-25
Last updated on: 2025-10-03
Assigner: MITRE
Description
Description
CVSS Scores
EPSS Scores
| Probability: | |
| Percentile: |
Meta Information
Affected Vendors & Products
| Vendor | Product | Version / Range |
|---|---|---|
| linuxfoundation | pytorch | From 2.6.0 (inc) to 2.7.0 (exc) |
Helpful Resources
Exploitability
| CWE ID | Description |
|---|---|
| CWE-1176 | The product performs CPU computations using algorithms that are not as efficient as they could be for the needs of the developer, i.e., the computations can be optimized further. |
Attack-Flow Graph
AI Powered Q&A
How can this vulnerability impact me? :
The impact of this vulnerability is that the dropout layers nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d may not behave as expected when fallback_random is true, potentially leading to inconsistent or incorrect model behavior during training or inference in PyTorch.
Can you explain this vulnerability to me?
This vulnerability involves the bernoulli_p decompose function in PyTorch versions before 3.7.0. The function lacks full consistency with the eager CPU implementation, which negatively affects the behavior of nn.Dropout1d, nn.Dropout2d, and nn.Dropout3d when fallback_random is set to true.