pytorch/benchmark

Pull Requests Commits

Update trainer.py

wconstab committed 3 years ago

Verified 11b51aa3

Minor fixes for distributed metrics

wconstab committed 3 years ago

4842d843

Fix backward-compatiblity issue when mps was not available in pytorch (#1129)

xuzhao9 committed 3 years ago

c7fd92c0

accept 2 outputs from parse_torchdynamo_args (#1128)

davidberard98 committed 3 years ago

1ec85e20

Also turn on JIT profiling mode (#1126)

davidberard98 committed 3 years ago

ba67ad70

Setup nvfuser automatic nightly CI (#1124)

xuzhao9 committed 3 years ago

031b028f

arg name fix (#1125)

zzpmiracle committed 3 years ago

e0392965

Add run contexts and dynamo to E2E model trainer (#1123)

xuzhao9 committed 3 years ago

7da92e9c

Add ability to set distributed to none for hf_bert e2e benchmark (#1121)

erichan1 committed 3 years ago

98a64dcc

Update the scribe category (#1120)

xuzhao9 committed 3 years ago

45f288d8

fix bug for opt_args.use_cosine_similarity (#1115)

FindHao committed 3 years ago

4808adea

Add gpu and cpu instances in scale-config.yaml (#1118)

xuzhao9 committed 3 years ago

077737ca

Add Dev Infra nightly test (#1114)

xuzhao9 committed 3 years ago

8c6114ec

added use_cosine_similarity as an extra argument (#1110)

snadampal committed 3 years ago

90ab3f9f

Prefetch data in timm_efficientdet (#1111)

xuzhao9 committed 3 years ago

ae17a00d

Remove the remaining fambench_dlrm directory (#1108)

xuzhao9 committed 3 years ago

17c1efff

Archive fambench_dlrm due to slow time-to-signal (#1107)

xuzhao9 committed 3 years ago

f3955b68

Enable CPU profiling when device='cuda' and profiling turned on (#1101)

davidberard98 committed 3 years ago

398b9713

Torchbench - explicitly enable JIT profiling executor (#1095)

davidberard98 committed 3 years ago

b83bd2be

add support for Execution Graph (#1077)

FindHao committed 3 years ago

6291bc09

Fix DLRM batch size argument to be mini_batch_size (#1082)

aaronenyeshi committed 3 years ago

9b9668fa

Enable flops for torchvision models (#748)

xuzhao9 committed 3 years ago

bf84f487

Update torchdynamo backend for torchbench (#1091)

davidberard98 committed 3 years ago

6d6d904b

shorten output format (#1073)

zzpmiracle committed 3 years ago

7ffb6e72

args.jit -> model.jit (#1086)

davidberard98 committed 3 years ago

97393f44

Handle models that don't have metadata.yaml (e.g. internal models) (#1085)

davidberard98 committed 3 years ago

d1fccd8b

seperate atol and rtol in correct check and read tol from os.env (#1072)

zzpmiracle committed 3 years ago

8a7d09d1

Fix eval batch size on Nvidia A100 40GB (#1078)

xuzhao9 committed 3 years ago

0e1a1885

Improve the release test by fixing bugs and add result analyzer (#1069)

xuzhao9 committed 3 years ago

8a7bbec9

Add release testing userbenchmark with multiple cuda versions (#1067)

xuzhao9 committed 3 years ago

3ffb2005

Older