preston-bernstein/ollama-yield

Stop your local LLM from holding the GPU when you start a game. Pauses and resumes Ollama around gaming and Plex transcodes. Works on any GPU, including AMD.

0 stars 0 forks 0 watchers Go MIT License

amd-gpu golang gpu homelab linux-gaming llm-inference local-llm ollama plex rocm self-hosted vram

View on GitHub

1 Open Issue Need Help Last updated: Sep 9, 2026

Open Issues Need Help

View All on GitHub

Preserve the prompt cache across a yield on llama.cpp backends 13 days ago

enhancement help wanted

preston-bernstein/ollama-yield

Stop your local LLM from holding the GPU when you start a game. Pauses and resumes Ollama around gaming and Plex transcodes. Works on any GPU, including AMD.

#amd-gpu#golang#gpu#homelab#linux-gaming#llm-inference#local-llm#ollama#plex#rocm#self-hosted#vram