How to run faster-whisper:gpu docker container with zluda to get AMD GPU usage

Open Teejer opened this issue 10 months ago • 1 comments

How would I use ZLUDA to launch the faster-whisper:gpu docker container so it uses my amd igpu? Specifically I'm running arch linux, but any general linux directions I'm sure I could adapt.

Thanks.

Jun 11 '25 18:06 Teejer

As far as I know, you can't. You can use the C2Translate version that works with ROCm and then run faster-whisper with it. You can look at my usage (via docker) here: https://github.com/Schaka/homeassistant-amd-pipeline

Jul 17 '25 08:07 Schaka