diff --git a/manifests/llama/rp.yaml b/manifests/llama/rp.yaml index 1273e34..c54b8e6 100644 --- a/manifests/llama/rp.yaml +++ b/manifests/llama/rp.yaml @@ -93,7 +93,7 @@ spec: # performance tuning - "--ctx-size" - - "32768" + - "16386" - "--parallel" - "1"