Merge branch 'v0.15.1-dev-fth' into 'v0.15.1-dev'

修复block-wise 用triton报错的bug

See merge request dcutoolkit/deeplearing/vllm!444

vllm/model_executor/layers/quantization/utils/fp8_utils.py

@@ -526,6 +526,7 @@ class W8A8BlockFp8LinearOp:
 
     def _run_triton(
         self,
+        out_features: int,
         input_2d: torch.Tensor,
         weight: torch.Tensor,
         weight_scale: torch.Tensor,