Skip to content

Add JetPack 7.1 support for NVIDIA Thor#1935

Draft
alexnorell wants to merge 2 commits intomainfrom
feature/jetpack-7.1-support
Draft

Add JetPack 7.1 support for NVIDIA Thor#1935
alexnorell wants to merge 2 commits intomainfrom
feature/jetpack-7.1-support

Conversation

@alexnorell
Copy link
Contributor

Summary

  • Adds Dockerfile for JetPack 7.1 (L4T R38.4.0) targeting NVIDIA Thor
  • Builds PyTorch and ONNX Runtime from source for CUDA 13.0 compatibility
  • Includes TensorRT 10.13 execution provider support

Test plan

  • Built and tested on Thor development kit
  • Verified PyTorch CUDA support
  • Verified ONNX Runtime with TensorRT and CUDA providers
  • Ran inference benchmarks

Adds Dockerfile.onnx.jetson.7.1.0 with support for:
- JetPack 7.1 / L4T R38.4.0
- CUDA 13.0 and cuDNN 9.12
- TensorRT 10.13
- PyTorch 2.10.0 (built from source)
- ONNX Runtime 1.25.0 with TensorRT EP (built from source)
- Flash Attention 2.8.3
- Python 3.12 on Ubuntu 24.04
@alexnorell alexnorell force-pushed the feature/jetpack-7.1-support branch from 78225d0 to 1b2f9cc Compare January 27, 2026 15:40
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants