Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. From 1.3.0 until 1.22.0, onnx.versionconverter.convertversion() can perform an out-of-bounds read in Gemm76::adaptgemm76() in onnx/versionconverter/adapters/gemm76.h when a Gemm node has input tensors with fewer than two dimensions because Bshape[1], Ashape[0], or A_shape[1] is accessed without a rank check, potentially causing a process crash during an opset 7 to 6 downgrade. This issue is fixed in version 1.22.0.