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

[PD Disaggregation] remove splitwise deployment on single node and refine the code (#4891)

Browse Source 
* remove splitwise deployment on single node and refine the code

* up

* up

* up

* add test

* up
This commit is contained in:
Juncai
2025-11-14 09:56:53 +08:00
committed by GitHub
parent 9703108c28
commit 36822fa49c
24 changed files with 626 additions and 963 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
fastdeploy/worker/worker_process.py
View File

@@ -929,8 +929,6 @@ def initialize_fd_config(args, ranks: int = 1, local_rank: int = 0) -> FDConfig:
logger.info(f"- Dynamic load weight: {load_config.dynamic_load_weight}")
logger.info(f"- Load strategy: {load_config.load_strategy}")
if args.splitwise_role != "mixed" and args.cache_transfer_protocol != "rdma":
envs.ENABLE_V1_KVCACHE_SCHEDULER = 0
if not current_platform.is_cuda() and not current_platform.is_xpu():
logger.info("Set ENABLE_V1_KVCACHE_SCHEDULER to 0 due to not supported.")
envs.ENABLE_V1_KVCACHE_SCHEDULER = 0