This commit is contained in:
2026-04-01 22:11:56 +08:00
parent 80c45e6749
commit edc6c5af84
+5 -5
View File
@@ -55,12 +55,12 @@
"local_path": "Qwen3.5-35B-A3B-GPTQ-Int4",
"dtype": "float16",
"quantization": "gptq",
"ctx": "65536",
"ctx": "262144",
"trust_remote": true,
"valid_tp": [1, 2],
"max_num_seqs": "4",
"max_num_seqs": "6",
"max_tokens": "4096",
"gpu_util": "0.85",
"gpu_util": "0.95",
"enforce_eager": false,
"tool_call_parser": "qwen3_coder",
"reasoning_parser": "qwen3",
@@ -87,11 +87,11 @@
},
"Qwen3-Next-80B-A3B-Instruct-AWQ-4bit": {
"local_path": "Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
"ctx": "65536",
"ctx": "262144",
"trust_remote": true,
"valid_tp": [2],
"max_num_seqs": "4",
"max_tokens": "8192",
"max_tokens": "4096",
"gpu_util": "0.90",
"enforce_eager": false,
"env": {