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

x

Browse Source
This commit is contained in:
chenjw28
2026-04-29 14:12:00 +08:00
parent 6e4e952e20
commit 5f85139bd4
1 changed files with 6 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
config.json
+6 -7
View File
@@ -69,19 +69,18 @@
"local_path": "Qwen3.6-27B-FP8", "local_path": "Qwen3.6-27B-FP8",
"dtype": "auto", "dtype": "auto",
"quantization": "fp8", "quantization": "fp8",
"ctx": "262144", "ctx": "32768",
"max_tokens": "65536", "max_tokens": "16384",
"max_num_batched_tokens": 32768, "max_num_batched_tokens": 8192,
"trust_remote": true, "trust_remote": true,
"enforce_eager": false, "enforce_eager": false,
"valid_tp": [ "valid_tp": [
1,
2 2
], ],
"max_num_seqs": 256, "max_num_seqs": 16,
"gpu_util": "0.92", "gpu_util": "0.80",
"kv_cache_dtype": "fp8", "kv_cache_dtype": "fp8",
"enable_prefix_caching": true, "enable_prefix_caching": false,
"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,