Is torch affected by CVE-2025-46153?

PyPI torch is affected in >=2.6.0 <2.7.0.

Is CVE-2025-46153 being exploited?

Not on the CISA KEV list. EPSS exploit probability is 0.4%.

medium

CVE-2025-46153

Q: Is CVE-2025-46153 being exploited?

Not on the CISA KEV list. EPSS exploit probability is 0.4%.

PyPI · torch

Summary

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.

Severity: medium
EPSS: 0.4% (p31)
Also known as: BIT-pytorch-2025-46153, PYSEC-2025-202
Published: 2025-09-25

References

Related advisories

CVE-2025-32434 — critical · PyPI/torch
CVE-2026-4538 — high · PyPI/torch
CVE-2025-55560 — high · PyPI/torch
CVE-2025-55558 — high · PyPI/torch
CVE-2025-55557 — high · PyPI/torch
CVE-2025-55553 — high · PyPI/torch
CVE-2025-55552 — high · PyPI/torch
CVE-2025-55551 — high · PyPI/torch

Is your project exposed to this? Stateward checks every dependency on every pull request and flags it only if your code actually reaches it.

Check my repo

Summarize with AI

ChatGPT Claude Perplexity

Sources: CISA KEV (public domain), OSV.dev & GitHub Advisory Database (CC-BY-4.0), FIRST EPSS, NVD/CWE (public domain). Served live from the Stateward advisory database.