add qwen3-vl thinking variant
This commit is contained in:
@@ -286,3 +286,22 @@ models:
|
||||
--presence-penalty 0.7
|
||||
--no-warmup
|
||||
--port ${PORT}
|
||||
|
||||
"Qwen3-VL-4B-Thinking-GGUF":
|
||||
ttl: 600
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3-VL-4B-Thinking-GGUF:Q4_K_M
|
||||
--n-gpu-layers 99
|
||||
--ctx-size 12288
|
||||
--predict 4096
|
||||
--flash-attn auto
|
||||
--jinja
|
||||
--top-p 0.95
|
||||
--top-k 20
|
||||
--temp 1.0
|
||||
--min-p 0.0
|
||||
--repeat-penalty 1.0
|
||||
--presence-penalty 0.0
|
||||
--no-warmup
|
||||
--port ${PORT}
|
||||
|
||||
Reference in New Issue
Block a user