暂不支持awq的scale和zero合并操作

2907adfa · zhuwenwen · d3fa2342 · 2907adfa
Commit 2907adfa authored Jan 06, 2026 by zhuwenwen
Hide whitespace changes
Inline Side-by-side

Showing with 2 additions and 2 deletions

vllm/model_executor/model_loader/utils.py vllm/model_executor/model_loader/utils.py +2 -2

No files found.
--- a/vllm/model_executor/model_loader/utils.py
+++ b/vllm/model_executor/model_loader/utils.py
@@ -252,8 +252,8 @@ def _get_model_architecture(model_config: ModelConfig) -> tuple[type[nn.Module],
                    
        # awq相关配置
        try:
-            if os.getenv('AWQ_MOE_SZ') == None:
-                os.environ['AWQ_MOE_SZ'] = '1'
+            # if os.getenv('AWQ_MOE_SZ') == None:
+            #     os.environ['AWQ_MOE_SZ'] = '1'
            if os.getenv('AWQ_PAD') == None and (torch.cuda.get_device_properties(torch.cuda.current_device()).multi_processor_count == 120):
                os.environ['AWQ_PAD'] = '1'
        except Exception as e: