Merge branch 'rx580_gpu' of https://github.com/mnccouk/ollama into rx580_gpu

This commit is contained in:
Matt 2024-09-19 21:11:45 +01:00
commit b0c75fd057

View file

@ -10,12 +10,12 @@ Get up and running with large language models.
### Linux with rx580 Radeon GPU ### Linux with rx580 Radeon GPU
This branch is has had changes for building on amd64 architecture(arm has been commented out in the Docker file) so as Ollama works with This branch is has had changes for building on amd64 architecture(arm has been commented out in the Docker file- just for build performance) so as Ollama works with
rx590 Redeon GPU. rx590 Redeon GPU.
It should be considered experimental. It should be considered experimental.
I've only been testing using the docker build. I've only been testing using the docker build, using ubuntu 22.04.04 LTS
Make sure docker is installed and running ok, and the docker host machine has rocm 5.7.1 libraries installed. Make sure docker is installed and running ok, and the docker host machine has rocm 5.7.1 libraries installed.