Immich uses ROCm to run GPU accelerated ML (facial recognition, smart search, etc). The current version (2.0.1) comes bundled with ROCm 6.4.3, which does not support RDNA 3.5 GPUs like the 890M. The first version that supports these GPUs is 6.4.4, so here are the steps to take when setting up Immich to be able to run ML models on the GPU.
- Make sure that your .env contains
HSA_OVERRIDE_GFX_VERSION=11.0.0 - You will need to extend the existing
immich_machine_learningimage and replace the ROCm runtime inside it.- Create a new
Dockerfilethat looks like thisFROM ghcr.io/immich-app/immich-machine-learning:v2.0.1-rocm
- Create a new