apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-17 14:11:14 +08:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

Fix block num in schduelr v1 for release 2.1 (#3315)

Browse Source 
* fix bug for scheduler v0

* fix block num setting in scheduler v1 for release 2.1

* fix block num setting in scheduler v1 for release 2.1

---------

Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
Co-authored-by: YUNSHEN XIE <1084314248@qq.com>
This commit is contained in:
chenjian
2025-08-12 00:41:05 +08:00
committed by GitHub
parent 9b07f85f6d
commit 25f51b0611
5 changed files with 33 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
fastdeploy/cache_manager/prefix_cache_manager.py
View File

@@ -64,7 +64,10 @@ class PrefixCacheManager:
self.speculative_config = config.speculative_config
self.local_data_parallel_id = local_data_parallel_id
self.num_gpu_blocks = self.cache_config.prefill_kvcache_block_num
if envs.ENABLE_V1_KVCACHE_SCHEDULER:
self.num_gpu_blocks = self.cache_config.total_block_num
else:
self.num_gpu_blocks = self.cache_config.prefill_kvcache_block_num
self.num_cpu_blocks = self.cache_config.num_cpu_blocks
self.gpu_free_block_list = list(range(self.num_gpu_blocks - 1, -1, -1))
if self.num_cpu_blocks > 0: