apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-09-27 21:02:24 +08:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

update flake8 version to support pre-commit in python3.12 (#3000)

Browse Source 
* update flake8 version to support pre-commit in python3.12

* polish code
This commit is contained in:
Zero Rains
2025-07-24 16:43:31 +08:00
committed by GitHub
parent 5151bc92c8
commit 0fb37ab7e4
30 changed files with 324 additions and 275 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
fastdeploy/model_executor/graph_optimization/cudagraph_piecewise_backend.py
View File

@@ -21,8 +21,8 @@ import paddle.nn.layer
from paddle.device.cuda import graphs
from fastdeploy.config import FDConfig
from fastdeploy.utils import get_logger
from fastdeploy.distributed.communication import capture_custom_allreduce
from fastdeploy.utils import get_logger
logger = get_logger("cudagrpah_piecewise_backend", "cudagraph_piecewise_backend.log")
@@ -99,7 +99,7 @@ class CudaGraphPiecewiseBackend:
entry.runnable(**kwargs)
logger.debug(
f"[CUDA GRAPH] Warm up for batch size {padding_batch_size}, "
f"finished ({n+1}/{entry.num_finished_warmup}) times"
f"finished ({n + 1}/{entry.num_finished_warmup}) times"
)
# Store input addresses for debug
@@ -114,7 +114,6 @@ class CudaGraphPiecewiseBackend:
new_grpah.capture_begin()
output = entry.runnable(**kwargs)
new_grpah.capture_end()
# Store output buffer
entry.cuda_graph = new_grpah