Snorkel MultitaskClassifier.load uses an unsafe torch.load

The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability CWE-502 in the MultitaskClassifier.load method of the MultitaskClassifier class. The method loads model weight files using torch.load without enabling the security-restrictive weightsonly=True parameter. This...

EUVD-2026-29502

The loadmodel function in the neuralmagictraining.py script of the optimate project in commit a6d302f912b481c94370811af6b11402f51d377f 2024-07-21 is vulnerable to insecure deserialization CWE-502. When loading a model state dictionary from a statedict.pt file via torch.load, the function does not...

PYSEC-2026-29

changedetection.io is a free open source web page change detection tool. In 0.54.9 and earlier, xpathfilter switches to XML mode for XML/RSS content and creates etree.XMLParserstripcdata=False without explicitly disabling external entity resolution, external DTD loading, or network-backed entity...

CVE-2026-41895

changedetection.io is a free open source web page change detection tool. In 0.54.9 and earlier, xpathfilter switches to XML mode for XML/RSS content and creates etree.XMLParserstripcdata=False without explicitly disabling external entity resolution, external DTD loading, or network-backed entity...

CVE-2026-31229

The Adversarial Robustness Toolbox ART thru 1.20.1 contains an insecure deserialization vulnerability CWE-502 in its Kubeflow component's model loading functionality. When loading model weights from a file e.g., model.pt during robustness evaluation, the code uses torch.load without the...

Deserialization of Untrusted Data

Overview lightning is a Deep Learning framework to train, deploy, and ship AI products Lightning fast. Affected versions of this package are vulnerable to Deserialization of Untrusted Data in the LightningModule.loadfromcheckpoint function. Any workflow that calls this function on an untrusted...

CVE-2026-31221

PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability CWE-502 in the checkpoint loading mechanism. The LightningModule.loadfromcheckpoint method, which is commonly used to load saved model states, internally calls torch.load without setting the...

CVE-2026-31224

The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability CWE-502 in the MultitaskClassifier.load method of the MultitaskClassifier class. The method loads model weight files using torch.load without enabling the security-restrictive weightsonly=True parameter. This...

The vulnerability in the backend/routers/files.py script of the system for launching and managing large language models in LoLLMS (Lord of Large Language Multimodal Systems) allows a hacker to write arbitrary files.

The vulnerability in the backend/routers/files.py script of the system for launching and managing large language models in LoLLMS Lord of Large Language Multimodal Systems involves unlimited loading of dangerous files. Exploiting this vulnerability allows a remote attacker to write any files they...

CVE-2026-31232: Deserialization of Untrusted Data

The CosyVoice project thru commit 6e01309e01bc93bbeb83bdd996b1182a81aaf11e 2025-30-21 contains an insecure deserialization vulnerability CWE-502 in its model loading process. When loading model files .pt from a user-specified directory via the --modeldir argument, the code uses torch.load without...

CVE-2026-31232

The CosyVoice project thru commit 6e01309e01bc93bbeb83bdd996b1182a81aaf11e 2025-30-21 contains an insecure deserialization vulnerability CWE-502 in its model loading process. When loading model files .pt from a user-specified directory via the --modeldir argument, the code uses torch.load without...

CVE-2026-31232

The CosyVoice project (up to commit 6e01309e01bc93bbeb83bdd996b1182a81aaf11e) is vulnerable to an insecure deserialization (CWE-502) flaw during its model loading process. The application uses torch.load() to process .pt model files from a user-specified directory (via the --model_dir argument) w...

CVE-2026-31237

The Ludwig framework (versions through 0.10.4 ) is vulnerable to insecure deserialization (CWE-502) within its predict() method. The issue occurs because the framework automatically identifies dataset file formats and uses pandas.read_pickle() to load .pkl files without any validation or security...

CVE-2026-31214

The torch-checkpoint-shrink.py script within the ml-engineering project (specifically commit 0099885db36a8f06556efe1faf552518852cb1e0) is vulnerable to insecure deserialization (CWE-502). The script utilizes torch.load() to process PyTorch checkpoint files (.pt) without enabling the security-rest...

CVE-2026-31232

The CosyVoice project thru commit 6e01309e01bc93bbeb83bdd996b1182a81aaf11e 2025-30-21 contains an insecure deserialization vulnerability CWE-502 in its model loading process. When loading model files .pt from a user-specified directory via the --modeldir argument, the code uses torch.load without...

PT-2026-40060

Name of the Vulnerable Software and Affected Versions PyTorch-Lightning versions 2.6.0 and earlier Description An insecure deserialization issue exists in the checkpoint loading mechanism. The load from checkpoint function in LightningModule internally calls torch.load without the weights only=Tr...

CVE-2026-31221

PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability CWE-502 in the checkpoint loading mechanism. The LightningModule.loadfromcheckpoint method, which is commonly used to load saved model states, internally calls torch.load without setting the...

CosyVoice 安全漏洞

CosyVoice is an open-source voice generation and AI voice cloning platform developed by FunAudioLLM. CosyVoice has a security vulnerability. This vulnerability arises from the model loading process, where the .pt files in the user-specified directory are loaded using torch.load, without enabling...

ludwig 安全漏洞

Ludwig is an open-source declarative deep learning framework developed by Ludwig. Versions of Ludwig 0.10.4 and earlier contain security vulnerabilities. These vulnerabilities stem from the model service component using torch.load without enabling the weightsonly=True parameter when loading model...

CVE-2026-31229

The Adversarial Robustness Toolbox ART thru 1.20.1 contains an insecure deserialization vulnerability CWE-502 in its Kubeflow component's model loading functionality. When loading model weights from a file e.g., model.pt during robustness evaluation, the code uses torch.load without the...