cjw/rocm_vllm_nightly
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Compare commits

base: cjw/rocm_vllm_nightly:7ff076530ace1e0b4f283a691eb1722f2432dadb
Branches Tags
cjw/rocm_vllm_nightly:main
..
compare: cjw/rocm_vllm_nightly:bccafdfa54ff2fbd5481ff057ffc21c087b6c7dd
Branches Tags
cjw/rocm_vllm_nightly:main

2 Commits
7ff076530a .. bccafdfa54

Author SHA1 Message Date
cjw bccafdfa54 x 2026-03-29 18:33:51 +08:00
cjw d84cdc00e5 x 2026-03-29 17:41:28 +08:00
1 changed files with 10 additions and 9 deletions
Expand all files Collapse all files
config.json
+10 -9
View File
@@ -24,7 +24,7 @@
"ctx": "24576",
"trust_remote": true,
"valid_tp": [2],
"max_num_seqs": "32",
"max_num_seqs": "8",
"max_tokens": "16384",
"gpu_util": "0.98",
"enforce_eager": false,
@@ -41,7 +41,7 @@
"ctx": "32768",
"trust_remote": true,
"valid_tp": [1, 2],
"max_num_seqs": "64",
"max_num_seqs": "8",
"max_tokens": "32768",
"gpu_util": "0.98",
"tool_call_parser": "qwen3_xml",
@@ -51,12 +51,13 @@
},
"Qwen3.5-27B-FP8": {
"local_path": "Qwen3.5-27B-FP8",
"dtype": "bfloat16",
"ctx": "65536",
"trust_remote": true,
"valid_tp": [1, 2],
"max_num_seqs": "64",
"max_tokens": "32768",
"gpu_util": "0.98",
"max_num_seqs": "4",
"max_tokens": "8192",
"gpu_util": "0.94",
"tool_call_parser": "qwen3_xml",
"enable_auto_tool_choice": true,
"served_model_name": "Qwen3.5-27B-FP8",
@@ -69,9 +70,9 @@
"ctx": "65536",
"trust_remote": true,
"valid_tp": [1, 2],
"max_num_seqs": "64",
"max_tokens": "32768",
"gpu_util": "0.98",
"max_num_seqs": "8",
"max_tokens": "4096",
"gpu_util": "0.92",
"tool_call_parser": "qwen3_xml",
"enable_auto_tool_choice": true,
"served_model_name": "Qwen3.5-35B-A3B-GPTQ-Int4",
@@ -83,7 +84,7 @@
"ctx": "65536",
"trust_remote": true,
"valid_tp": [1, 2],
"max_num_seqs": "64",
"max_num_seqs": "8",
"max_tokens": "32768",
"gpu_util": "0.98",
"tool_call_parser": "qwen3_xml",