user/settings/llama/swap.yaml (view raw)
1 2 3 4 5 6 7 8 9 10 | # yaml-language-server: $schema=https://raw.githubusercontent.com/mostlygeek/llama-swap/refs/heads/main/config-schema.json
models:
gpt-oss:20b:
cmd: llama-server --port ${PORT} -hf ggml-org/gpt-oss-20b-GGUF --jinja -ub 2048 -b 2048
gpt-oss:120b:
cmd: llama-server --port ${PORT} -hf ggml-org/gpt-oss-120b-GGUF --jinja -ub 2048 -b 2048
hooks:
on_startup:
preload:
- "gpt-oss:120b"
|