CVE-2025-46153
published 2025-09-25CVE-2025-46153: 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…
PriorityP426medium5.3CVSS 3.1
AVNACLPRNUINSUCLINAN
EPSS
0.39%
31.3th percentile
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.
Affected
4 ranges
| Vendor | Product | Version range | Fixed in |
|---|---|---|---|
| debian | pytorch | — | — |
| linuxfoundation | pytorch | >= 2.6.0 < 2.7.0 | 2.7.0 |
| msrc | azl3_pytorch_2.2.2-7_on_azure_linux_3.0 | — | — |
| msrc | cbl2_pytorch_2.0.0-9_on_cbl_mariner_2.0 | — | — |
CVSS provenance
nvdv3.15.3MEDIUMCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N
osv5.3MEDIUM
vendor_debian5.3MEDIUM
vendor_msrc5.3MEDIUM
vendor_redhat5.3MEDIUM
Stop checking back — get the weekly exploitation signal.
Every Monday: what got weaponized or added to CISA KEV in the last seven days — each CVE cross-linked to its PoC, Nuclei template, and detection rule. Free, one email a week, unsubscribe in one click.
Red Hat
torch: PyTorch inconsistent results
vendor_redhat·2025-09-25·CVSS 5.3
CVE-2025-46153 [MEDIUM] CWE-440 torch: PyTorch inconsistent results
torch: PyTorch inconsistent results
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.
A consistency flaw has been discovered in the PyTorch library. PyTorch 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.
Mitigation: Mitigation for this issue is either not available or the currently available options do not meet the Red Hat Product Security criteria comprising ease of use and deployment, applicability to widespread installat
Microsoft
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,
vendor_msrc·2025-09-09·CVSS 5.3
CVE-2025-46153 [MEDIUM] CWE-1176 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,
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.
FAQ: Is Azure Linux the only Microsoft product that includes this open-source library and is therefore potentially affected by this vulnerability?
One of the main benefits to our customers who choose to use the Azure Linux distro is the commitment to keep it up to date with the most recent and most secure versions of the open source libraries with which the distro is composed. Microsoft is committed to transparency in this work which is why we began publishing CSAF/VEX in October 2025. See this blog post for more information. If impact to additional
Debian
CVE-2025-46153: pytorch - PyTorch before 3.7.0 has a bernoulli_p decompose function in decompositions.py e...
vendor_debian·2025·CVSS 5.3
CVE-2025-46153 [MEDIUM] CVE-2025-46153: pytorch - PyTorch before 3.7.0 has a bernoulli_p decompose function in decompositions.py e...
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.
Scope: local
bookworm: open
bullseye: open
forky: open
sid: open
trixie: open
GHSA
GHSA-h36m-9qjr-pv93: PyTorch before 3
ghsa_unreviewed·2025-09-25
CVE-2025-46153 [MEDIUM] CWE-1176 GHSA-h36m-9qjr-pv93: PyTorch before 3
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.
OSV
CVE-2025-46153: PyTorch before 3
osv·2025-09-25·CVSS 5.3
CVE-2025-46153 [MEDIUM] CVE-2025-46153: PyTorch before 3
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.
No detection rules found.
No public exploits indexed.
No writeups or analysis indexed.
2025-09-25
Published