hardware accelerated machine learning and transcoding using AMD 9070XT in WSL2
Trying to get ML and transcoding working via ROCm and VAAPI-WSL. Running Immich v1.135.3
ROCm has been installed per https://rocm.docs.amd.com/projects/radeon/en/latest/docs/install/wsl/install-radeon.html#
WSL kernel has been updated to 6.11.9 - GPU is accessible in ubuntu through WSL2.
latest compose/hwaccel/hwtranscode .ymls used - docker compose up fails if devices: - /dev/kfd:/dev/kfd is left in - changing to /dev/dxg:/dev/dxg results in everything running happily but an error showing up in the docker log when you try to run machine learning and similar errors when transcoding.
error in next post due to character limit
ROCm has been installed per https://rocm.docs.amd.com/projects/radeon/en/latest/docs/install/wsl/install-radeon.html#
WSL kernel has been updated to 6.11.9 - GPU is accessible in ubuntu through WSL2.
latest compose/hwaccel/hwtranscode .ymls used - docker compose up fails if devices: - /dev/kfd:/dev/kfd is left in - changing to /dev/dxg:/dev/dxg results in everything running happily but an error showing up in the docker log when you try to run machine learning and similar errors when transcoding.
error in next post due to character limit
