pytorch/benchmark

Pull Requests Commits

Improve the compute score script to accept json directory or json file without requiring a config. (#144) (#145)

xuzhao9 committed 5 years ago

Verified 98934653

Ensure input to benchmark upload is a list of messages (#139)

wconstab committed 5 years ago

Verified 86a5e8f8

Clarify README w.r.t. test.py (#134)

wconstab committed 5 years ago

Verified dccfd6f5

Wconstab/score (#131)

wconstab committed 5 years ago

Verified 9537dc4c

Upload raw benchmark data, up to a max of 100 samples per model (#129)

wconstab committed 5 years ago

Verified c91eeed5

Include instructions for benchmark initial setup (#133)

wconstab committed 5 years ago

Verified 83b04ee5

Disable intermittently failing maskrcnn eval (#132)

wconstab committed 5 years ago

Verified 65c5ffe2

Merge pull request #130 from pytorch/nikitha_removeDLRM

nikithamalgifb committed 5 years ago

Verified db516ab6

Remove duplicated DLRM model from the top level benchmark repo

nikithamalgifb committed 5 years ago

999f6d6b

Include date in benchmark filename (#122)

wconstab committed 5 years ago

Verified 6296022f

Add logic to push benchmark json to github to archive (#110)

wconstab committed 5 years ago

Verified 9f432c2a

fix a broken link to ADDING_MODELS.md (#120)

Krovatkin committed 5 years ago

Verified ece76e2a

Add stargan model (#119)

EdwardTyantov committed 5 years ago

Verified a77854fb

Add DLRM Model (#112)

nikithamalgifb committed 5 years ago

Verified 06ab5e8c

Avoid git operation triggered in yolo model (#118)

wconstab committed 5 years ago

Verified 5a077ba8

Empty cache after each test, gc before each measurement (#117)

wconstab committed 5 years ago

Verified 3525d687

cleanup workdir, and document new model structure

zdevito committed 5 years ago

cc244c36

Remove accidental model in top level folder and fix name in correct place (#113)

wconstab committed 5 years ago

Verified 000566b8

Move torch version check to test json and update repo fields (#97)

wconstab committed 5 years ago

Verified 60fec86b

Avoid extremely slow cyclegan on PR jobs to speed up feedback, but still run on nightly (#102)

wconstab committed 5 years ago

Verified 512c94e8

Add maskrcnn_benchmark

wconstab committed 5 years ago

Verified cdf3a6ce

Merge pull request #76 from pytorch/wconstab/nits

wconstab committed 5 years ago

Verified 9369b711

Make device and jit support in tacotron model more explicit

wconstab committed 5 years ago

4fcd14af

Merge pull request #105 from pytorch/wconstab/typo

wconstab committed 5 years ago

Verified 6691e903

make this work without changing directories

zdevito committed 5 years ago

5f1f52b0

Get rid of path manipulation

zdevito committed 5 years ago

5768afa0

Merge pull request #101 from pytorch/wconstab/slow

wconstab committed 5 years ago

Verified 98e80cbc

Fix spelling of attention.. dir name

wconstab committed 5 years ago

e4530e08

Disable cpu training of models taking many seconds per iteration

wconstab committed 5 years ago

faf0138d

Model names are valid python identifiers

zdevito committed 5 years ago

6ddfe568

Newer Older