Skip to content

Commit eb57a00

Browse files
author
niushengxiao
committed
fix: fix cache length
1 parent 814d267 commit eb57a00

1 file changed

Lines changed: 1 addition & 1 deletion

File tree

lightllm/server/httpserver/manager.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -704,7 +704,7 @@ async def _wait_to_token_package(
704704
prompt_cache_len = metadata.pop("prompt_cache_len", 0)
705705
cpu_prompt_cache_len = metadata.pop("cpu_prompt_cache_len", 0)
706706
disk_prompt_cache_len = metadata.pop("disk_prompt_cache_len", 0)
707-
metadata["prompt_cache_len"] = prompt_cache_len
707+
metadata["prompt_cache_len"] = prompt_cache_len + cpu_prompt_cache_len + disk_prompt_cache_len
708708
sub_req_id_to_mtp_accepted_token_num[sub_req_id] = metadata.get("mtp_accepted_token_num", 0)
709709

710710
if is_first_token:

0 commit comments

Comments
 (0)