minor cleanup

LucasWilkinson · LucasWilkinson · commit 4f301fa76476 · 2025-07-15T22:35:53.000-04:00
Signed-off-by: Lucas Wilkinson &lt;lwilkins@redhat.com&gt;
diff --git a/vllm/v1/attention/backends/mla/common.py b/vllm/v1/attention/backends/mla/common.py
@@ -477,7 +477,7 @@ def __init__(self,
             self.cudnn_workspace = torch.empty(
                 CUDNN_WORKSPACE_SIZE * scheduler_config.max_num_seqs,
                 dtype=torch.int8,
-                device=self.device,
+                device=device,
             )
 
     def _build_fi_prefill_wrappers(self, prefill: FlashInferPrefillMetadata):

Original file line number	Diff line number	Diff line change
`@@ -477,7 +477,7 @@ def __init__(self,`
`477`	`477`	`self.cudnn_workspace = torch.empty(`
`478`	`478`	`CUDNN_WORKSPACE_SIZE * scheduler_config.max_num_seqs,`
`479`	`479`	`dtype=torch.int8,`
`480`		`- device=self.device,`
	`480`	`+ device=device,`
`481`	`481`	`)`
`482`	`482`
`483`	`483`	`def _build_fi_prefill_wrappers(self, prefill: FlashInferPrefillMetadata):`