[PD] Add different TP sizes support for no-MLA models (#6793)

Co-authored-by: shangmingc <csmthu@gmail.com>
Co-authored-by: Shangming Cai <caishangming@linux.alibaba.com>
This commit is contained in:
Hongbo Xu
2025-06-25 02:00:22 -07:00
committed by GitHub
co-authored by shangmingc Shangming Cai
parent f3cbd24541
commit e21aa1df67
3 changed files with 265 additions and 27 deletions
@@ -27,6 +27,8 @@ class KVArgs:
decode_tp_size: int
# for pp prefill
prefill_pp_size: int
kv_head_num: int
page_size: int
class KVPoll: