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 14, 05:30 UTC
[Bugfix][CI] Fix `test_remote_decode_lifecycle.py::test_short_prompt_lifecycle` (#22727)
Signed-off-by: NickLucche <nlucches@redhat.com>
Nicolò Lucchesi committed
Aug 12, 2025 at 16:30 UTC
3d9d40efdeea7011dc3c496ad9d55cfdc90aff92
1 file changed
+3
-2
tests/v1/kv_connector/unit/test_remote_decode_lifecycle.py
+3
-2
@@ -130,8 +130,9 @@ def test_short_prompt_lifecycle():
130
# Confirm we do not have any memory leaks after req lifecycle.
131
# We need to mark sending finish to clear data for persistent batch.
132
scheduler_output = scheduler.schedule()
133
- model_runner_output = copy.deepcopy(EMPTY_MODEL_RUNNER_OUTPUT)
134
- model_runner_output.finished_sending = [request.request_id]
133
+ # Use create_model_runner_output to pass kv_connector_output along
134
+ model_runner_output = create_model_runner_output(
135
+ reqs=[request], finished_sending=[request.request_id])
136
scheduler.update_from_output(scheduler_output, model_runner_output)
137
assert_scheduler_empty(scheduler)
138