search

microsoft / onnxruntime

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
https://onnxruntime.ai
MIT License
14.81k stars 2.94k forks source link

[Build] [On-device-training] Running onnxruntime-training (CPU) Python Demo on Debian 12 aarch64 Environment #22270

Open Leo5050xvjf opened 1 month ago

Leo5050xvjf commented 1 month ago

Describe the issue

Env:

Ram

Description: Hello,

  1. I am currently working with the onnxruntime-training demo, which I've successfully run on my PC and on Colab using Python, achieving satisfactory results. Now, I wish to deploy this project on my development board, a Rockchip RK3576. However, I've encountered an issue where I can't directly install the necessary packages via:

pip install onnxruntime pip install onnxruntime-training

It appears that PyPI does not offer pre-compiled packages for my development board's environment. After some research, I found the following pre-compiled packages at this release page: https://github.com/Microsoft/onnxruntime/releases/tag/v1.19.2. I believe the files I need are:

Is my understanding correct?

  1. As I am new to this, I'm unsure how to use these .tgz files with Python after downloading. Can the project only be invoked via C/C++, or is there a way to utilize these with Python?

  2. Can I perform On-Device Training (CPU Only) using Python or C/C++ on Debian 12, arm64? Are there any example codes for ONNX Runtime training in C/C++? I have not found any relevant information so far, only Python code.

Thank you! I look forward to your response, thank you!

Urgency

Yes, this issue is quite urgent for me as it affects my current work progress. I hope to receive a response as soon as possible so that I can continue advancing my project. Thank you for your understanding and support!

Target platform

aarch64

Build script

None

Error / output

None

Visual Studio Version

No response

GCC / Compiler Version

No response

github-actions[bot] commented 2 weeks ago

This issue has been automatically marked as stale due to inactivity and will be closed in 30 days if no further activity occurs. If further support is needed, please provide an update and/or more details.