apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-12-24 13:28:13 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Revert "[Optim] Remove limitation of number of kvcache blocks (#5612)" (#5702)

Browse Source 
This reverts commit 9da89a374b.
This commit is contained in:
Divano
2025-12-23 15:41:33 +08:00
committed by GitHub
parent 0bef9b684f
commit c1aa66df02
7 changed files with 12 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/usage/environment_variables.md
View File

@@ -88,9 +88,5 @@ environment_variables: dict[str, Callable[[], Any]] = {
# Count for cache_transfer_manager process error
"FD_CACHE_PROC_ERROR_COUNT": lambda: int(os.getenv("FD_CACHE_PROC_ERROR_COUNT", "10")),
# Max allocated KV cache blocks. Use this to limit how many KV cache blocks the engine is allowed to allocate.
# Set to -1 (default) for no limit, or a positive integer to cap the maximum number of blocks that can be allocated.
"FD_MAX_KVCACHE_BLOCKS": lambda: int(os.getenv("FD_MAX_KVCACHE_BLOCKS", "-1")),
}
```

3
docs/zh/usage/environment_variables.md
View File

@@ -88,7 +88,4 @@ environment_variables: dict[str, Callable[[], Any]] = {
# cache_transfer_manager 进程残留时连续错误阈值
"FD_CACHE_PROC_ERROR_COUNT": lambda: int(os.getenv("FD_CACHE_PROC_ERROR_COUNT", "10")),}
# KVCache Block块分配值的上限。此变量限制引擎分配的块数上限。当为默认值-1时表示不设限
"FD_MAX_KVCACHE_BLOCKS": lambda: int(os.getenv("FD_MAX_KVCACHE_BLOCKS", "-1")),
```