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 5, 09:45 UTC
[CI Bug] Fix pre-commit issue in main (#39347)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Apr 8, 2026 at 17:10 UTC
3352bf8b03ce29610f2ea0ec7d6855b6160c2d01
1 file changed
+4
-1
tests/v1/e2e/spec_decode/test_spec_decode.py
+4
-1
@@ -214,7 +214,10 @@ def test_ngram_gpu_default_with_async_scheduling(
214
async_scheduling=async_scheduling,
215
)
216
# Assert the resolved async_scheduling config matches what was requested.
217
- assert spec_llm.llm_engine.vllm_config.scheduler_config.async_scheduling == async_scheduling
217
+ assert (
218
+ spec_llm.llm_engine.vllm_config.scheduler_config.async_scheduling
219
+ == async_scheduling
220
+ )
221
evaluate_llm_for_gsm8k(spec_llm, expected_accuracy_threshold=0.8)
222
del spec_llm
223
cleanup_dist_env_and_memory()