pytorch/benchmark

Pull Requests Commits

Krovatkin committed 4 years ago

d695d142

Add time_lazy.py script for lazy tensor measurement

wconstab committed 4 years ago

4fbf86a1

Make bert_pytorch support lazy device

wconstab committed 4 years ago

6f602558

Add check_lazy.py to sweep models and collect stats

wconstab committed 4 years ago

31e733d3

Delete models that I couldn't install (NOT FOR MERGE)

wconstab committed 4 years ago

fb1ff187

[LTC][Not for merge] Enabling Lazy Tensor in hf_Bert and resnet50

Jiewen Tan committed 4 years ago

65a44fb6

Fix the import issue of speech_transformer. (#542)

xuzhao9 committed 4 years ago

bdea82c6

Fix GPU utilization issue of mobilenet_v3_large (#543)

xuzhao9 committed 4 years ago

1b2fd45c

Fix GPU utilization issue of timm_vovnet and timm_vision_transformer (#540)

xuzhao9 committed 4 years ago

e6ef04b4

Fixes regnet and resnest models (#535)

xuzhao9 committed 4 years ago

107cd55d

Improve DLRM utilization on CUDA (#534)

aaronenyeshi committed 4 years ago

30fb7339

Upload bisection result to GitHub artifact. (#525)

xuzhao9 committed 4 years ago

0fdd791d

Set the same warmup rounds for profiler and non-profiler. (#531)

xuzhao9 committed 4 years ago

e27e0c92

Initial TorchBench Summary Metadata Validation (#526)

aaronenyeshi committed 4 years ago

91d11f1c

Add backward-compatible layer to `torch.jit.script()` API change (#522)

xuzhao9 committed 4 years ago

107328ec

Fix fastNLP GPU utilization issue (#499)

xuzhao9 committed 4 years ago

7f40d308

Fix a few bugs in v1 nightly workflow. (#505)

xuzhao9 committed 4 years ago

6b53cc3f

Add missing AWS region env to workflows (#510)

driazati committed 4 years ago

917cbd9e

Remove the fastNLP model. (#506)

xuzhao9 committed 4 years ago

ee651e18

Enable writing stats to metrics.pytorch.org (#501)

driazati committed 4 years ago

47fb9de5

Improve the GPU utilization for timm_nfnet and timm_efficientnet (#502)

xuzhao9 committed 4 years ago

afa744e1

Automatically detect anomaly, report and bisect it if detected (#488)

xuzhao9 committed 4 years ago

ad061568

Increase DenseNet121 Batch Size for better Utilization (#496)

aaronenyeshi committed 4 years ago

8c83aa2b

Increase the time limit for bisection job. (#500)

xuzhao9 committed 4 years ago

6a83be61

Collect CPU time using time_ns in run.py (#497)

aaronenyeshi committed 4 years ago

19c590c2

Improve the sweep plotting script to use V1 score (#480)

xuzhao9 committed 4 years ago

0ed0d872

Small fixes to run.py (#495)

aaronenyeshi committed 4 years ago

bde07fc0

Update CITATION.cff (#491)

orionr committed 4 years ago

0fa26bbe

Fix run.py --profile options to support choosing device activity (#493)

aaronenyeshi committed 4 years ago

2d2142c8

Upgrade the environment to "Pendulum-v1". (#492)

xuzhao9 committed 4 years ago

05da0886

Older