use llama-swap peer config instead of load balancing
1 file changed, 9 insertions(+), 0 deletions(-)
changed files
M user/settings/llama/swap-config.yml → user/settings/llama/swap-config.yml
@@ -65,3 +65,12 @@ hooks: on_startup: preload: - sweep-next-edit-v2-7b + +peers: + prefect: + proxy: http://prefect:11343 + models: + - sweep-next-edit + - sweep-next-edit-v2 + - zeta20 + - zeta21