chore(llama-swap): tune presence penalty to 1.5 and remove repeat penalty
This commit is contained in:
@@ -44,8 +44,7 @@ in
|
|||||||
--top-p 0.95 \
|
--top-p 0.95 \
|
||||||
--top-k 20 \
|
--top-k 20 \
|
||||||
--min-p 0.0 \
|
--min-p 0.0 \
|
||||||
--presence-penalty 0.0 \
|
--presence-penalty 1.5 \
|
||||||
--repeat-penalty 1.0 \
|
|
||||||
-dev CUDA0 \
|
-dev CUDA0 \
|
||||||
-fit off \
|
-fit off \
|
||||||
--chat-template-kwargs "{\"preserve_thinking\": true}"
|
--chat-template-kwargs "{\"preserve_thinking\": true}"
|
||||||
|
|||||||
Reference in New Issue
Block a user