[Build] ONNX Run Time on Conda Forge - Add CUDA Support

[RoyiAvital]

Describe the issue

I installed ONNX Run Time using conda install onnxruntime -c conda-forge.

I run onnxruntime.get_available_providers() and get only the CPU as executer (CPUExecutionProvider).

My environment has PyTorch installed and CUDA working (Just as evidence that CUDA is well defined and working on my system).

Looking at the conda logs I can see it installs onnxruntime-1.20.1-py312h414cfab_0_cpu.

Could you please provide CUDA support through the conda-forge package?

Urgency

No response

Target platform

Windows 11 + CUDA (NVIDIA GPU)

Build script

conda install onnxruntime -c conda-forge

Error / output

Linking onnxruntime-1.20.1-py312h414cfab_0_cpu

Visual Studio Version

No response

GCC / Compiler Version

No response

[snnn]

The package is not maintained by us. You may raise the issue to conda-forge's issue list.

[snnn]

https://github.com/conda-forge/onnxruntime-feedstock

[RoyiAvital]

Thanks for the pointer.