7 matches found
ONNX: Out-of-bounds Read In ConvTranspose Shape Inference Via Undersized Weight Rank
Summary ONNX ConvTranspose shape inference performed an out-of-bounds read when the weight tensor had fewer than three dimensions and the model did not provide a "kernelshape" attribute. The inference function derived an empty "effectivekernelshape" vector from the undersized weight tensor, then...
ONNX: Out-of-bounds read in ConvTranspose shape inference via undersized weight rank
Summary ONNX ConvTranspose shape inference performed an out-of-bounds read when the weight tensor had fewer than three dimensions and the model did not provide a kernelshape attribute. The inference function derived an empty effectivekernelshape vector from the undersized weight tensor, then...
GHSA-6PWM-X34P-FVJ2 ONNX: Out-of-bounds read in ConvTranspose shape inference via undersized weight rank
Summary ONNX ConvTranspose shape inference performed an out-of-bounds read when the weight tensor had fewer than three dimensions and the model did not provide a kernelshape attribute. The inference function derived an empty effectivekernelshape vector from the undersized weight tensor, then...
CVE-2026-115920
A flaw was found in ONNX. When processing a crafted machine learning model containing a transpose convolution ConvTranspose operator with an undersized weight tensor, an out-of-bounds read occurs during dimension calculation. An attacker can exploit this issue by supplying a malicious model to an...
AZL-112169 CVE-2026-115920 affecting package pytorch 2.2.2-17
Open Neural Network Exchange ONNX is an open standard for machine learning interoperability. From 1.2.1 until 1.22.0, ConvTranspose shape inference processes a weight tensor with fewer than three dimensions without a kernelshape attribute by deriving an empty effectivekernelshape and indexing it...
CVE-2026-115920 ONNX: Out-of-bounds read in ConvTranspose shape inference via undersized weight rank
Open Neural Network Exchange ONNX is an open standard for machine learning interoperability. From 1.2.1 until 1.22.0, ConvTranspose shape inference processes a weight tensor with fewer than three dimensions without a kernelshape attribute by deriving an empty effectivekernelshape and indexing it...
Open Neural Network Exchange ONNX buffer error vulnerability
Open Neural Network Exchange version 1.2.1 through 1.22.0 contains a buffer error vulnerability. This vulnerability arises when ConvTranspose shape inference processes weight tensors with fewer than three dimensions and no kernelshape attribute; it results in an empty effectivekernelshape being...