改用Dflash格式推理
This commit is contained in:
+5
-4
@@ -78,16 +78,17 @@
|
|||||||
1,
|
1,
|
||||||
2
|
2
|
||||||
],
|
],
|
||||||
"max_num_seqs": 256,
|
"max_num_seqs": 32,
|
||||||
"gpu_util": "0.92",
|
"gpu_util": "0.90",
|
||||||
"kv_cache_dtype": "fp8",
|
|
||||||
"enable_prefix_caching": true,
|
|
||||||
"tool_call_parser": "qwen3_coder",
|
"tool_call_parser": "qwen3_coder",
|
||||||
"reasoning_parser": "qwen3",
|
"reasoning_parser": "qwen3",
|
||||||
"enable_auto_tool_choice": true,
|
"enable_auto_tool_choice": true,
|
||||||
"language_model_only": true,
|
"language_model_only": true,
|
||||||
"served_model_name": "Qwen3.6-27B-FP8",
|
"served_model_name": "Qwen3.6-27B-FP8",
|
||||||
"hf_model_id": "Qwen/Qwen3.6-27B-FP8",
|
"hf_model_id": "Qwen/Qwen3.6-27B-FP8",
|
||||||
|
"env": {
|
||||||
|
"VLLM_RPC_TIMEOUT": "300"
|
||||||
|
},
|
||||||
"speculative": {
|
"speculative": {
|
||||||
"method": "dflash",
|
"method": "dflash",
|
||||||
"model": "Qwen3.6-27B-DFlash",
|
"model": "Qwen3.6-27B-DFlash",
|
||||||
|
|||||||
Reference in New Issue
Block a user