pytorch/benchmark

Pull Requests Commits

fix name and assignment for device_id

FindHao committed 2 years ago

6136b1b7

set numba logger level to ERROR

FindHao committed 2 years ago

89a48a9b

FindHao committed 2 years ago

54934822

update logger levels and messages

FindHao committed 2 years ago

0d143417

comment warning

FindHao committed 2 years ago

46aa78d5

fix compatibility for pynvml

FindHao committed 2 years ago

f64d62cd

add a probe round

FindHao committed 2 years ago

3a2b61c5

run more iterations for short models

FindHao committed 2 years ago

24b35547

update error message

FindHao committed 2 years ago

07e30a63

FindHao committed 2 years ago

f005777a

replace nvidia-ml-py with pynvml;fix return value type hints

FindHao committed 2 years ago

06e8859d

FindHao committed 3 years ago

1231cbb7

FindHao committed 3 years ago

57f0649d

disable warning messages; add gpu device id to output

FindHao committed 3 years ago

8c208556

make it compatible with older pynvml

FindHao committed 3 years ago

dc09b5f3

fix bug for gpu detection

FindHao committed 3 years ago

832bafab

move memory usage detection to metrics.py

FindHao committed 3 years ago

43c58619

remove print log for metrics

FindHao committed 3 years ago

7fb59251

add dependent nvidia-ml-py

FindHao committed 3 years ago

7cc42b99

enable peak memory measurement by default

FindHao committed 3 years ago

5fcdd4c9

Add the --mps_only option. (#1432)

kulinseth committed 3 years ago

1635a749

Disable the nvfuser nightly CI (#1431)

xuzhao9 committed 3 years ago

4fb6838c

Fix torchrec_dlrm model and enable fbcode test (#1428)

xuzhao9 committed 3 years ago

08fc5914

Use the correct batch size for timm ViT to improve stableness (#1419)

xuzhao9 committed 3 years ago

06538fd1

Upgrade to gcc-11 in the A100 Docker to fix CI error (#1427)

xuzhao9 committed 3 years ago

e2dfe592

Add torchtext to fix benchmark installation error (#1424)

spzala committed 3 years ago

0e6a7de9

Upgrade the actions used by the bisection workflow (#1417)

xuzhao9 committed 3 years ago

d2c692c2

replace torchbench impelmentation with torchrec that shards the module (#1401)

colin2328 committed 3 years ago

9806a7ec

Fix vision model on inductor and profile on CPU only machines (#1415)

xuzhao9 committed 3 years ago

7bd50b01

change fastNLP_Bert's task name (#1410)

FindHao committed 3 years ago

3fe6446f

Older