microsoft / onnxruntime

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
https://onnxruntime.ai
MIT License
14.57k stars 2.92k forks source link

[Documentation] How Configure CUDA 12.* and cuDNN for GPU with ONNX Runtime and C# on Windows 11 #21212

Open GeorgeS2019 opened 4 months ago

GeorgeS2019 commented 4 months ago

Describe the documentation issue

Current information limited to Cuda 11. Need Update and How to make it works with Cuda 12.

https://onnxruntime.ai/docs/tutorials/csharp/csharp-gpu.html

Page / URL

https://onnxruntime.ai/docs/tutorials/csharp/csharp-gpu.html

GeorgeS2019 commented 4 months ago

For example https://nietras.com/2024/04/28/phi-3-mini-csharp-ortgenai/

CUDA is required for the cuda models, which I’ll use here, and you need to have a compatible GPU and drivers installed.

Where can we have the updated information for Cuda 12.*

GeorgeS2019 commented 4 months ago

The documentation is only up to onnxruntime 1.16

Is 1.18.1 compatible with Cuda 12.*

GeorgeS2019 commented 4 months ago

we also need this how to make it work for windows using c# codes

https://github.com/microsoft/onnxruntime/blob/gh-pages/docs/genai/tutorials/phi3-v.md

GeorgeS2019 commented 4 months ago

https://github.com/microsoft/onnxruntime/blob/gh-pages/docs/tutorials/on-device-training/index.md

Missing other c# examples

GeorgeS2019 commented 4 months ago

C# (future)

We need codes for c# future https://github.com/microsoft/onnxruntime/blob/gh-pages/docs/execution-providers/CUDA-ExecutionProvider.md

GeorgeS2019 commented 4 months ago

In this page https://github.com/microsoft/onnxruntime/blob/gh-pages/docs/tutorials/csharp/fasterrcnn_csharp.md

Download the Faster R-CNN ONNX model to your local system.

FasterRCNN-10.onnx <= deleted

Right link https://github.com/onnx/models/tree/main/validated/vision/object_detection_segmentation/faster-rcnn/model

github-actions[bot] commented 2 months 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.

GeorgeS2019 commented 2 months ago

@sophies927

So far, it is not clear if there has been improvement in the documentation.