[Documentation] Latest Supported Onnx Runtime for TensorRT mentioned as 12.2

[Samiii777]

Describe the documentation issue

Not sure if I am correct or not, but I realized onnx runtime for tensorrt recently has been added for CUDA 12.2 in here:
docs/execution-providers/TensorRT-ExecutionProvider.md
https://github.com/microsoft/onnxruntime/blob/gh-pages/docs/execution-providers/CUDA-ExecutionProvider.md#requirements

according to Nvidia's documentation TensorRT 8.6 GA is only supported by CUDA 12.1 and not 12.2
https://developer.nvidia.com/nvidia-tensorrt-8x-download
am I missing something ?

Page / URL

https://github.com/microsoft/onnxruntime/blob/gh-pages/docs/execution-providers/CUDA-ExecutionProvider.md#requirements

[tianleiwu]

Because of CUDA Minor Version Compatibility, Onnx Runtime built with CUDA 12.2 shall be compatible with CUDA 12.1 and 12.3.

For example, Nvidia docker nvcr.io/nvidia/pytorch:24.01-py3 has CUDA 12.3 and TRT 8.6 GA. They work well with Onnx Runtime 1.17 in my tests.

[Samiii777]

thanks for clarifying