Skip to content

P
Paddle

PaddlePaddle / Paddle
大约 1 年 前同步成功

通知 2298
Star 20931
Fork 5422
P
Paddle
已关闭
Opened by saxon_zh@saxon_zhOwner

Results differ when doing CPU inference multiple times with MKLDNN

Created by: cryoco

System information Paddle version: 1.8.2/1.8.3 Paddle With CUDA: False OS: Ubuntu 16.04 CPU: 16 Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz Python version: 3.5.2 CUDA version: 9.0.176 cuDNN version: None.None.None Nvidia driver version: None API information: inference configuration

config.disable_gpu()
config.enable_mkldnn()
config.set_cpu_math_library_num_threads(4)

To Reproduce

  1. download models
  1. perform inference with mkldnn code to reproduce: test_ocr_mkldnn_diff.txt

This issue can be reproduced with any input. We use data with all ones here.

python3 test_ocr_mkldnn_diff.py --model_file=./ch_det_mv3_db/model --params_file=./ch_det_mv3_db/params --mkldnn

Run inference 10 times with the same model and input, and result shifts every time(only the first result is correct). image

  1. perform inference without mkldnn
python3 test_ocr_mkldnn_diff.py --model_file=./ch_det_mv3_db/model --params_file=./ch_det_mv3_db/params

The results are consistent. image