mirrors/PaddleClas
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/PaddleClas.git synced 2025-06-03 21:55:06 +08:00
Code Issues Projects Releases Wiki Activity

Fix static training speed (#1590)

Browse Source 
* fix training speed

* update config setting method
This commit is contained in:
Wei Shengyu 2021-12-23 11:13:51 +08:00 committed by sibo2rr
parent 7eec82b87e
commit bce0d041a7
2 changed files with 7 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
ppcls/configs/ImageNet/ResNet/ResNet50_fp16.yaml
View File

@ -16,6 +16,7 @@ Global:
save_inference_dir: ./inference save_inference_dir: ./inference
# training model under @to_static # training model under @to_static
to_static: False to_static: False
use_dali: True
# mixed precision training # mixed precision training
AMP: AMP:

13
ppcls/static/train.py
View File

@ -81,14 +81,13 @@ def main(args):
# amp related config # amp related config
if 'AMP' in config: if 'AMP' in config:
AMP_RELATED_FLAGS_SETTING = { AMP_RELATED_FLAGS_SETTING = {
'FLAGS_cudnn_exhaustive_search': "1", 'FLAGS_cudnn_exhaustive_search': 1,
'FLAGS_conv_workspace_size_limit': "1500", 'FLAGS_conv_workspace_size_limit': 1500,
'FLAGS_cudnn_batchnorm_spatial_persistent': "1", 'FLAGS_cudnn_batchnorm_spatial_persistent': 1,
'FLAGS_max_indevice_grad_add': "8", 'FLAGS_max_inplace_grad_add': 8,
"FLAGS_cudnn_batchnorm_spatial_persistent": "1",
} }
for k in AMP_RELATED_FLAGS_SETTING: os.environ['FLAGS_cudnn_batchnorm_spatial_persistent'] = '1'
os.environ[k] = AMP_RELATED_FLAGS_SETTING[k] paddle.fluid.set_flags(AMP_RELATED_FLAGS_SETTING)
use_xpu = global_config.get("use_xpu", False) use_xpu = global_config.get("use_xpu", False)
use_npu = global_config.get("use_npu", False) use_npu = global_config.get("use_npu", False)