Commit Graph

48 Commits (dac0679a70e88dba467e4c7539e74cc8da513a7c)

Author SHA1 Message Date
Luo Tao 7d19b65e65 move paddle/scripts/check_env.sh to benchmark/
7 years ago
qingqing01 24509f4af9 Fix the grammar in copyright. (#8403)
7 years ago
dzhwinter e983cc90fc
"fix decode bug" (#7711)
7 years ago
dzhwinter b9b75377a2
Feature/hooks (#7513)
7 years ago
ranqiu92 e8483dde8b
Merge pull request #7311 from ranqiu92/plot
7 years ago
zhangchao 893a15f5e2
Update plotlog.py
7 years ago
zhangchao 452caa0954
Update plotlog.py
7 years ago
zhangchao cd9374fba2
Update plotlog.py
7 years ago
zhangchao 71337d20ea
Update the plotlog.py
7 years ago
ranqiu f59ad3cee4 Add copyright notice to plotlog.py
7 years ago
ranqiu 207642657e Add script to plot learning curve
7 years ago
Luo Tao 24920c32eb update openblas benchmark when export OPENBLAS_MAIN_FREE=1
7 years ago
tensor-tang d14ca1c39f fix inference crash of alexnet benchmark
7 years ago
tensor-tang d630d39214 auto set openblas env when inference and remove unused env for openblas
7 years ago
tensor-tang cf9e09b115 set openblas env to avoid threads conflicts
7 years ago
Tao Luo 658dbe6c98
Merge pull request #6852 from tensor-tang/alexnet
7 years ago
tensor-tang 025a6f3c23 unify the test reorder
7 years ago
tensor-tang a785496b69 fix logical error
7 years ago
tensor-tang 81e15bcf23 reduce the training samples for infer model
7 years ago
tensor-tang 6fc454486c reduce test_period to save time when training openblas
7 years ago
tensor-tang 86b8bdc0af enable inference alexnet benchmark
7 years ago
tensor-tang 19939657ea enable training alexnet benchmark
7 years ago
tensor-tang ee49f54e7f use small samples to infer openblas for saving time.
7 years ago
Luo Tao b0e4357178 seperate openblas train/infer script with mkl
7 years ago
Luo Tao a524f8e6d6 auto compute the training FPS by avg elapsed time
7 years ago
tensor-tang 7b827d95ad use awk command to replace bc
7 years ago
tensor-tang 79b17097f6 cal FPS of inference result
7 years ago
tensor-tang aef639448c skip train list when inference, skip test list when training
7 years ago
tensor-tang a5aac61410 skip cost when inference
7 years ago
tensor-tang 849bf9d0d0 separate mkldnn benchmark as train and infer
7 years ago
tensor-tang 6c5f928a3e enable inference benchmark
7 years ago
tensor-tang 74dca2733d enable mkldnn benchmark with googlenet
7 years ago
tensor-tang a3b2b7b1c7 remove the hard code setting in benchmark scripts
7 years ago
tensor-tang b007055e9d reduce the lr in case of nan in small batchsize
7 years ago
tensor-tang 2dff98ca11 remove auto setting from HT, since it's hard to unify with MacOS
7 years ago
tensor-tang 714fa9e37c remove some topology tests
7 years ago
tensor-tang 30b57eef40 auto KMP setting with HT
7 years ago
tensor-tang 2f3665e988 update reset script for benchmark
7 years ago
tensor-tang cfa86a3f70 should reset env every time
7 years ago
tensor-tang 83310a1c14 refine mkldnn benchmark script
7 years ago
tensor-tang 799f80ad33 remove useless code
7 years ago
tensor-tang 86a9434c71 rename script and modify comments
7 years ago
tensor-tang d72f636c2e add KMP setting and default test vgg19
7 years ago
tensor-tang 0f7c40713f add vgg and script for mkldnn benchmark
8 years ago
dangqingqing 88a7cbd639 Use empty file list for paddle/image.
8 years ago
dangqingqing b1cc9da4b7 Merge branch 'develop' of https://github.com/baidu/Paddle into benchmark_cfg_doc
8 years ago
dangqingqing 9d377f0996 Update doc and merge run_multi.sh into run.sh for PaddlePaddle.
8 years ago
dangqingqing a8342d073d Add benchmark config and document
8 years ago