shiyuan680 3b6b3c4214 [MODELRUNNERV2]fix penality ops (#7013) ...

### What this PR does / why we need it?
fix penality ops for new version, and achieved a 10% performance
improvement

### How was this patch tested?
pytest
‎tests/e2e/nightly/single_node/ops/singlecard_ops/triton/test_penality.py
- vLLM version: v0.16.0
- vLLM main:
15d76f74e2

Signed-off-by: shiyuan680 <917935075@qq.com>

2026-03-11 17:13:34 +08:00

..

__init__.py

[Feature] support eager mode in model runner v2 (#5210)

2025-12-29 15:28:34 +08:00

gumbel.py

[Feature] adapt to uva buffer and main2main (#6657)

2026-02-12 10:36:31 +08:00

penalties.py

[MODELRUNNERV2]fix penality ops (#7013)

2026-03-11 17:13:34 +08:00

sampler.py

[Feature] adapt to uva buffer and main2main (#6657)

2026-02-12 10:36:31 +08:00