x

2026-03-29 00:29:54 +08:00
parent e186ee4136
commit 55d5a4c416
1 changed files with 4 additions and 4 deletions
@@ -29,12 +29,12 @@
      "hf_model_id": "RedHatAI/Qwen3-14B-FP8-dynamic"
    },
    "Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit": {
-      "ctx": "24576",
+      "ctx": "131072",  // 128K
      "trust_remote": true,
      "valid_tp": [1, 2],
-      "max_num_seqs": "64",
-      "max_tokens": "32768",
-      "gpu_util": "0.98",
+      "max_num_seqs": "32",  // 适当减少以节省显存
+      "max_tokens": "65536",
+      "gpu_util": "0.95",  // 适当降低以预留更多显存
      "hf_model_id": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit"
    },
    "Qwen3-Next-80B-A3B-Instruct-AWQ-4bit": {