vllm-project/vllm-ascend

[Misc]: Discussion on accuracy variance

Chiusa

#6623 aperta il 9 feb 2026

 (2 commenti) (0 reazioni) (0 assegnatari)C++ (2048 fork)github user discovery
help wantedstalewait-feedback

Metriche repository

Star
 (2637 stelle)
Metriche merge PR
 (Merge medio 4g 5h) (559 PR mergiate in 30 g)

Descrizione

Anything you want to discuss about vllm on ascend.

Because of batch variance, We cannot guarantee that the same input will yield the same output in a multi-batch inference case. And this accuracy variance is more explicit in certain datasets.

model dataset acc acc variance
deepseek-v3.1 GPQA 74~82 8
qwen3-235b GPQA 64~71 7
qwen3-480b GPQA 60~67 7

We need to test these case on GPU or use other inference engine, such as sglang to check if this is an acc bug.

If GPU also has the same acc variance like us, we believe that this acc variance is reasonable for this dataset. Otherwise, we need to solve this acc bug.

Guida contributor