Skip to content

Commit 6bfc96c

Browse files
Muuuchenjimoosciuc
authored andcommitted
[FIX] Fix concatenation error in capture_bs when open --disable-cuda-graph-padding and without MTP (sgl-project#5412)
1 parent 097c8d6 commit 6bfc96c

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

python/sglang/srt/model_executor/cuda_graph_runner.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -117,7 +117,7 @@ def get_batch_sizes_to_capture(model_runner: ModelRunner):
117117
if capture_bs is None:
118118
if server_args.speculative_algorithm is None:
119119
if server_args.disable_cuda_graph_padding:
120-
capture_bs = list(range(1, 33)) + range(40, 161, 16)
120+
capture_bs = list(range(1, 33)) + list(range(40, 161, 16))
121121
else:
122122
capture_bs = [1, 2, 4, 8] + list(range(16, 161, 8))
123123
else:

0 commit comments

Comments
 (0)