Pytorch

Don't miss out!

Thousands of developers use stack.watch to stay informed.
Get an email whenever new security vulnerabilities are reported in any Pytorch product.

RSS Feeds for Pytorch security vulnerabilities

Create a CVE RSS feed including security vulnerabilities found in Pytorch products with stack.watch. Just hit watch, then grab your custom RSS feed url.

Products by Pytorch Sorted by Most Security Vulnerabilities since 2018

Pytorch22 vulnerabilities

Pytorch Torchserve2 vulnerabilities

By the Year

In 2026 there have been 2 vulnerabilities in Pytorch with an average score of 7.1 out of ten. Last year, in 2025 Pytorch had 16 security vulnerabilities published. Right now, Pytorch is on track to have less security vulnerabilities in 2026 than it did last year. However, the average CVE base score of the vulnerabilities in 2026 is greater by 1.19.

Year	Vulnerabilities	Average Score
2026	2	7.05
2025	16	5.86
2024	3	0.00
2023	2	7.55
2022	1	9.80

It may take a day or so for new Pytorch vulnerabilities to show up in the stats or in the list of recent security vulnerabilities. Additionally vulnerabilities may be tagged under a different product or component name.

Recent Pytorch Security Vulnerabilities

CVE	Date	Vulnerability	Products
CVE-2026-4538	Mar 22, 2026	PyTorch 2.10.0 pt2 Loading Handler deserialization flaw local only A vulnerability was identified in PyTorch 2.10.0. The affected element is an unknown function of the component pt2 Loading Handler. The manipulation leads to deserialization. The attack can only be performed from a local environment. The exploit is publicly available and might be used. The project was informed of the problem early through a pull request but has not reacted yet.	Pytorch
CVE-2026-24747	Jan 27, 2026	PyTorch weights_only unpickler memory corruption risk AOE before v2.10.0 PyTorch is a Python package that provides tensor computation. Prior to version 2.10.0, a vulnerability in PyTorch's `weights_only` unpickler allows an attacker to craft a malicious checkpoint file (`.pth`) that, when loaded with `torch.load(..., weights_only=True)`, can corrupt memory and potentially lead to arbitrary code execution. Version 2.10.0 fixes the issue.	Pytorch
CVE-2025-46152	Sep 25, 2025	PyTorch<2.7.0 bitwise_right_shift OOB Value Bug In PyTorch before 2.7.0, bitwise_right_shift produces incorrect output for certain out-of-bounds values of the "other" argument.	Pytorch
CVE-2025-46148	Sep 25, 2025	PyTorch 2.6.0 nn.PairwiseDistance(p=2) incorrect results In PyTorch through 2.6.0, when eager is used, nn.PairwiseDistance(p=2) produces incorrect results.	Pytorch
CVE-2025-46149	Sep 25, 2025	PyTorch <=2.6.0 nn.Fold Assertion Error via Inductor In PyTorch before 2.7.0, when inductor is used, nn.Fold has an assertion error.	Pytorch
CVE-2025-46150	Sep 25, 2025	PyTorch <2.7.0 FMP2D Inconsistent Results with torch.compile In PyTorch before 2.7.0, when torch.compile is used, FractionalMaxPool2d has inconsistent results.	Pytorch
CVE-2025-46153	Sep 25, 2025	PyTorch <3.7.0 bernoulli_p decompose causes dropout1d/2d/3d loss 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.	Pytorch
CVE-2025-55551	Sep 25, 2025	Pytorch 2.8.0 torch.linalg.lu DoS via Slice Ops An issue in the component torch.linalg.lu of pytorch v2.8.0 allows attackers to cause a Denial of Service (DoS) when performing a slice operation.	Pytorch
CVE-2025-55552	Sep 25, 2025	PyTorch 2.8.0 Unexpected Behavior with torch.rot90 + torch.randn_like pytorch v2.8.0 was discovered to display unexpected behavior when the components torch.rot90 and torch.randn_like are used together.	Pytorch
CVE-2025-55553	Sep 25, 2025	PyTorch 2.7.0 Syntax Error in proxy_tensor.py Causes DoS A syntax error in the component proxy_tensor.py of pytorch v2.7.0 allows attackers to cause a Denial of Service (DoS).	Pytorch