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

[Graph Optimization][BugFix][CI] Fix 0size bug && add unitest (#5495)

Browse Source
This commit is contained in:
Ryan
2025-12-11 16:25:26 +08:00
committed by GitHub
parent e1347be4d9
commit e58fed3665
7 changed files with 25 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tests/ce/deploy/ernie45t_21b_cinn.yaml Normal file
View File

@@ -0,0 +1,8 @@
max_model_len: 32768
max_num_seqs: 128
tensor_parallel_size: 1
quantization: wint4
graph_optimization_config:
graph_opt_level: 2
sot_warmup_sizes: [2,16,32,64]
use_cudagraph: True

2
tests/e2e/test_EB_VL_Lite_sot_serving.py
View File

@@ -91,7 +91,7 @@ def setup_and_run_server():
"--reasoning-parser",
"ernie-45-vl",
"--graph-optimization-config",
'{"graph_opt_level": 1, "use_cudagraph": true, "full_cuda_graph": false}',
'{"graph_opt_level": 2, "use_cudagraph": true, "full_cuda_graph": false}',
]
# Start subprocess in new process group