GitLab

Signed-off-by: ZX-ModelCloud <zx@modelcloud.ai>
Co-authored-by: ZX-ModelCloud <zx@modelcloud.ai>

Support penalty in overlap mode; return logprob with chunked prefill; improve benchmark scripts (#3988)
Co-authored-by: SangBin Cho <rkooo567@gmail.com>
Co-authored-by: dhou-xai <dhou@x.ai>
Co-authored-by: Hanming Lu <hanming_lu@berkeley.edu>

Co-authored-by: SangBin Cho <rkooo567@gmail.com>

Co-authored-by: Chayenne <zhaochen20@outlook.com>
Co-authored-by: yizhang2077 <1109276519@qq.com>

Co-authored-by: Zhangyi <1109276519@qq.com>

Co-authored-by: dhou-xai <dhou@x.ai>

Co-authored-by: yukavio <kavioyu@gmail.com>

Co-authored-by: Yang Zheng(SW)(Alex) <you@example.com>

Fix illegal memory access in overlap mode & Use more fused triton kernels for building meta data (#2051)

Co-authored-by: Yineng Zhang <me@zhyncs.com>

Co-authored-by: Ying Sheng <sqy1415@gmail.com>
Co-authored-by: Yineng Zhang <me@zhyncs.com>
Co-authored-by: Lianmin Zheng <lianminzheng@gmail.com>
Co-authored-by: Liangsheng Yin <hnyls2002@gmail.com>