xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-20 20:12:29 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Core] latency optimization (#3890)

Browse Source
This commit is contained in:
youkaichao 2024-04-06 19:14:06 -07:00 committed by GitHub
parent 95baec828f
commit 2f19283549
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/core/block_manager_v1.py
View File

@ -328,7 +328,7 @@ class BlockSpaceManagerV1(BlockSpaceManager):
self,
seq: Sequence,
) -> bool:
token_ids_len = len(seq.data.get_token_ids())
token_ids_len = seq.data.get_len()
return token_ids_len > 0 and token_ids_len % seq.block_size == 0
def _maybe_promote_last_block(