microsoft / microsoft/onnxruntime
[Feature Request] Jetson Thor Cuda 13.0 Onnxruntime-gpu
- Dominant language
- C++
- Stars
- 21.9k
- Forks
- 4.2k
- Avg merge
- 4d 11h
- Merged PRs (30d)
- 184
Description
Can be added aarch64 gpu version for jetson cuda 13 devices?
Hi,
I am currently using Jetson Thor for my project (CUDA 13). Some of NGC containers are back compatible with Thor but TensorRT-25.06. It's too hard to combine different packages.
I checked https://aiinfra.visualstudio.com/PublicPackages/_artifacts/feed/ort-cuda-13-nightly/PyPI/onnxruntime-gpu/overview/1.24.0.dev20260120001 and i couldn't see aarch64.
Contributor guide
Research direction
Start with the linked ort-cuda-13 nightly PyPI artifact page and inspect the repository's build and packaging entry points for the existing CUDA 13 artifacts. Done means an aarch64 onnxruntime-gpu package is available for Jetson Thor and CUDA 13, with the relevant packaging checks passing.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- build-system, machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 42/100