apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-08 01:50:27 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Optimize] Increase zmq buffer size to prevent apiserver too slowly to consume (#3723) (#3728)

Browse Source 
Co-authored-by: chenjian <1435317881@qq.com>
This commit is contained in:
RAM
2025-08-30 15:58:18 +08:00
committed by GitHub
parent 64cf769bee
commit b9af800edd
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
fastdeploy/envs.py
View File

@@ -44,7 +44,7 @@ environment_variables: dict[str, Callable[[], Any]] = {
# Whether to use HuggingFace tokenizer. # Whether to use HuggingFace tokenizer.
"FD_USE_HF_TOKENIZER": lambda: os.getenv("FD_USE_HF_TOKENIZER", 0), "FD_USE_HF_TOKENIZER": lambda: os.getenv("FD_USE_HF_TOKENIZER", 0),
# Set the high watermark (HWM) for receiving data during ZMQ initialization # Set the high watermark (HWM) for receiving data during ZMQ initialization
"FD_ZMQ_SNDHWM": lambda: os.getenv("FD_ZMQ_SNDHWM", 10000), "FD_ZMQ_SNDHWM": lambda: os.getenv("FD_ZMQ_SNDHWM", 64000),
# cache kv quant params directory # cache kv quant params directory
"FD_CACHE_PARAMS": lambda: os.getenv("FD_CACHE_PARAMS", "none"), "FD_CACHE_PARAMS": lambda: os.getenv("FD_CACHE_PARAMS", "none"),
# Set attention backend. "NATIVE_ATTN", "APPEND_ATTN" # Set attention backend. "NATIVE_ATTN", "APPEND_ATTN"