GPU support takes 4.0 more GB

This commit is contained in:
David PAGNON 2024-07-26 21:48:20 +02:00 committed by GitHub
parent 16c8a73bbf
commit 5eb6ef0fcc
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -128,7 +128,7 @@ If you don't use Anaconda, type `python -V` in terminal to make sure python>=3.9
``` ```
4. ***Optional:***\ 4. ***Optional:***\
*For faster inference, you can run on the GPU. Install pyTorch with CUDA and cuDNN support, and ONNX Runtime with GPU support (not available on MacOS).* *For faster inference, you can run on the GPU. Install pyTorch with CUDA and cuDNN support, and ONNX Runtime with GPU support (not available on MacOS). However, be aware that it takes an additional 4 Go on disk.*
Go to the [pyTorch website]( https://pytorch.org/get-started/locally), select the latest CUDA version that is also [available with ONNX runtime](https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), and run the provided command.\ Go to the [pyTorch website]( https://pytorch.org/get-started/locally), select the latest CUDA version that is also [available with ONNX runtime](https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), and run the provided command.\
For example, for Windows 11 (June 6th, 2024), CUDA 12.4 is not available for pyTorch, and CUDA 12.1 is not available for ONNX Runtime, so you should revert to CUDA 11.8: For example, for Windows 11 (June 6th, 2024), CUDA 12.4 is not available for pyTorch, and CUDA 12.1 is not available for ONNX Runtime, so you should revert to CUDA 11.8: