Compare commits
2 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
| bccafdfa54 | |||
| d84cdc00e5 |
+10
-9
@@ -24,7 +24,7 @@
|
|||||||
"ctx": "24576",
|
"ctx": "24576",
|
||||||
"trust_remote": true,
|
"trust_remote": true,
|
||||||
"valid_tp": [2],
|
"valid_tp": [2],
|
||||||
"max_num_seqs": "32",
|
"max_num_seqs": "8",
|
||||||
"max_tokens": "16384",
|
"max_tokens": "16384",
|
||||||
"gpu_util": "0.98",
|
"gpu_util": "0.98",
|
||||||
"enforce_eager": false,
|
"enforce_eager": false,
|
||||||
@@ -41,7 +41,7 @@
|
|||||||
"ctx": "32768",
|
"ctx": "32768",
|
||||||
"trust_remote": true,
|
"trust_remote": true,
|
||||||
"valid_tp": [1, 2],
|
"valid_tp": [1, 2],
|
||||||
"max_num_seqs": "64",
|
"max_num_seqs": "8",
|
||||||
"max_tokens": "32768",
|
"max_tokens": "32768",
|
||||||
"gpu_util": "0.98",
|
"gpu_util": "0.98",
|
||||||
"tool_call_parser": "qwen3_xml",
|
"tool_call_parser": "qwen3_xml",
|
||||||
@@ -51,12 +51,13 @@
|
|||||||
},
|
},
|
||||||
"Qwen3.5-27B-FP8": {
|
"Qwen3.5-27B-FP8": {
|
||||||
"local_path": "Qwen3.5-27B-FP8",
|
"local_path": "Qwen3.5-27B-FP8",
|
||||||
|
"dtype": "bfloat16",
|
||||||
"ctx": "65536",
|
"ctx": "65536",
|
||||||
"trust_remote": true,
|
"trust_remote": true,
|
||||||
"valid_tp": [1, 2],
|
"valid_tp": [1, 2],
|
||||||
"max_num_seqs": "64",
|
"max_num_seqs": "4",
|
||||||
"max_tokens": "32768",
|
"max_tokens": "8192",
|
||||||
"gpu_util": "0.98",
|
"gpu_util": "0.94",
|
||||||
"tool_call_parser": "qwen3_xml",
|
"tool_call_parser": "qwen3_xml",
|
||||||
"enable_auto_tool_choice": true,
|
"enable_auto_tool_choice": true,
|
||||||
"served_model_name": "Qwen3.5-27B-FP8",
|
"served_model_name": "Qwen3.5-27B-FP8",
|
||||||
@@ -69,9 +70,9 @@
|
|||||||
"ctx": "65536",
|
"ctx": "65536",
|
||||||
"trust_remote": true,
|
"trust_remote": true,
|
||||||
"valid_tp": [1, 2],
|
"valid_tp": [1, 2],
|
||||||
"max_num_seqs": "64",
|
"max_num_seqs": "8",
|
||||||
"max_tokens": "32768",
|
"max_tokens": "4096",
|
||||||
"gpu_util": "0.98",
|
"gpu_util": "0.92",
|
||||||
"tool_call_parser": "qwen3_xml",
|
"tool_call_parser": "qwen3_xml",
|
||||||
"enable_auto_tool_choice": true,
|
"enable_auto_tool_choice": true,
|
||||||
"served_model_name": "Qwen3.5-35B-A3B-GPTQ-Int4",
|
"served_model_name": "Qwen3.5-35B-A3B-GPTQ-Int4",
|
||||||
@@ -83,7 +84,7 @@
|
|||||||
"ctx": "65536",
|
"ctx": "65536",
|
||||||
"trust_remote": true,
|
"trust_remote": true,
|
||||||
"valid_tp": [1, 2],
|
"valid_tp": [1, 2],
|
||||||
"max_num_seqs": "64",
|
"max_num_seqs": "8",
|
||||||
"max_tokens": "32768",
|
"max_tokens": "32768",
|
||||||
"gpu_util": "0.98",
|
"gpu_util": "0.98",
|
||||||
"tool_call_parser": "qwen3_xml",
|
"tool_call_parser": "qwen3_xml",
|
||||||
|
|||||||
Reference in New Issue
Block a user