A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 12, 03:30 UTC
[BugFix] Fix assert in `build_for_cudagraph_capture` (#29893)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Lucas Wilkinson committed
Dec 2, 2025 at 19:56 UTC
5cdd66450910589c8e1a3d25e80711b0b6e51eb1
1 file changed
+1
-1
vllm/v1/worker/gpu_model_runner.py
+1
-1
@@ -4000,7 +4000,7 @@ class GPUModelRunner(
4000
num_reqs=num_reqs_padded,
4001
max_query_len=max_query_len,
4002
ubatch_slices=ubatch_slices,
4003
- for_cudagraph_capture=True,
4003
+ for_cudagraph_capture=is_graph_capturing,
4004
)
4005
4006
with self.maybe_dummy_run_with_lora(