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

[BugFix] fix vl performance bug (#5181)

Browse Source 
* fix vl performance bug

* update code

* update code

---------

Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
This commit is contained in:
kevin
2025-11-26 21:06:52 +08:00
committed by GitHub
parent 209970836e
commit bf30f45738
2 changed files with 10 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
fastdeploy/engine/common_engine.py
View File

@@ -691,8 +691,16 @@ class EngineService:
else:
max_num_batched_tokens = self.cfg.model_config.max_model_len
# In multi-mode scenarios, using available_block_num to pull requests to prevent heavy rescheduling
# in the frequency domain due to insufficient blocks
if self.cfg.model_config.enable_mm:
self.resource_manager.check_and_free_block_tables()
available_blocks = self.resource_manager.available_block_num()
else:
available_blocks = self.cfg.cache_config.max_block_num_per_seq
tasks = self.scheduler.get_requests(
available_blocks=self.cfg.cache_config.max_block_num_per_seq,
available_blocks=available_blocks,
block_size=self.cfg.cache_config.block_size,
reserved_output_blocks=self.cfg.cache_config.enc_dec_block_num,
max_num_batched_tokens=max_num_batched_tokens,