apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 08:37:06 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Excutor] Experiment Feature-Support Prefill in cudagraph (#3459)

Browse Source 
* Support prefill in Cudagraph

* Refactor GetBlockShapeAndSplitKVBlock Kernel V2

* Refactor GetBlockShapeAndSplitKVBlock Kernel V2.1

* Refactor GetBlockShapeAndSplitKVBlock Kernel V2.2

* Refactor GetBlockShapeAndSplitKVBlock Kernel V2.3

* Refactor GetBlockShapeAndSplitKVBlock Kernel V2.4

* Refactor GetBlockShapeAndSplitKVBlock Kernel V2.5

* Solve problem about encoder_num_blocks_x_cpu

* Add early-exit mechanism for attention kernel

* fix test case about append-attention

* Update testcode, Add annotations to related tensors

* move get_input_length_list

* solve test_code

* Add annotations about early-exit for attention kernel

* Add annotations about early-exit for attention kernel2

* solve comment

* solve mtp

---------

Co-authored-by: RAM <gstian5555@outlook.com>
This commit is contained in:
Jundong Liu
2025-09-08 13:12:24 +08:00
committed by GitHub
parent 472402bf4e
commit 3d0aaa5923
21 changed files with 528 additions and 260 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
fastdeploy/worker/worker_process.py
View File

@@ -397,7 +397,6 @@ class PaddleDisWorkerProc:
self.get_profile_block_num_signal.value[0] = num_blocks_local
else:
num_blocks_local = self.fd_config.parallel_config.total_block_num
logger.info(f"------- num_blocks_global: {num_blocks_local} --------")
# wait engine launch cache_manager
if self.cache_config.enable_prefix_caching or self.parallel_config.splitwise_role != "mixed":