Skip to content

Commit 4f877ff

Browse files
committed
fix: change num_tokens_across_dp type from int to torch.Tensor
Signed-off-by: Jade Zheng <zheng.shoujian@outlook.com>
1 parent c6780fc commit 4f877ff

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

vllm_ascend/worker/model_runner_v1.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1817,7 +1817,7 @@ def _dummy_run(
18171817
is_compile: bool = False,
18181818
with_prefill: bool = True,
18191819
skip_attn: bool = True,
1820-
num_tokens_across_dp: Optional[int] = None,
1820+
num_tokens_across_dp: Optional[torch.Tensor] = None,
18211821
) -> torch.Tensor:
18221822
# Set num_scheduled_tokens based on num_tokens and max_num_seqs
18231823
# for dummy run with LoRA so that the num_reqs collectively

0 commit comments

Comments
 (0)