PYSEC-2025-202

PyTorch before 3.7.0 has a bernoullip 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 fallbackrandom=True...

CVE-2025-46153

The PyTorch machine learning framework (versions before 3.7.0 ) contains a vulnerability in the bernoulli_p decompose function located in decompositions.py . The issue stems from a lack of full consistency with the eager CPU implementation, which negatively impacts nn.Dropout1d , nn.Dropout2d , a...

PT-2025-39383

Name of the Vulnerable Software and Affected Versions PyTorch versions prior to 3.7.0 Description The software contains an inconsistency in the bernoulli p decompose function within decompositions.py. This function does not fully align with the eager CPU implementation, which impacts the...