Commits

Branch selector

 master

User selector

rwightman

Datepicker

All time

Commit History

Commits on Aug 19, 2021

• git push origin masterMerge branch 'yohann84L-fix_accuracy'

  rwightman

  committed
  acd6c68

  Copy full SHA for acd6c68

• Tweak accuracy topk safety. Fix #807

  rwightman

  committed
  d667351

  Copy full SHA for d667351

• Add lamb/lars to optim init imports, remove stray comment

  rwightman

  committed
  a16a753

  Copy full SHA for a16a753

• MOAR optimizer changes. Woo!

  rwightman

  committed
  c207e02

  Copy full SHA for c207e02

• Fix lars tests

  rwightman

  committed
  42c1f0c

  Copy full SHA for 42c1f0c

• More optimizer cleanup. Change all to no longer use .data. Improve (b)float16 use with adabelief. Add XLA compatible Lars.

  rwightman

  committed
  a426511

  Copy full SHA for a426511

Commits on Aug 18, 2021

• One more scalar -> tensor fix for lamb optimizer

  rwightman

  committed
  9541f49

  Copy full SHA for 9541f49

• Update lamp.py comment

  rwightman

  authored
  8f68193

  Copy full SHA for 8f68193

• Merge pull request #813 from rwightman/opt_cleanup

  Show description for 4d28401
  rwightman

  authored
  4d28401

  Copy full SHA for 4d28401

• add madgradw optimizer

  rwightman

  committed
  a6af48b

  Copy full SHA for a6af48b

• Remove experiment from lamb impl

  rwightman

  committed
  55fb5ee

  Copy full SHA for 55fb5ee

• A few optimizer comments, dead import, missing import

  rwightman

  committed
  8a9eca5

  Copy full SHA for 8a9eca5

• Add optimizer tests and update testing to pytorch 1.9

  rwightman

  committed
  959eaff

  Copy full SHA for 959eaff

• Optimizer improvements, additions, cleanup

  Show description for ac469b5
  rwightman

  committed
  ac469b5

  Copy full SHA for ac469b5

Commits on Aug 15, 2021

• Merge pull request #805 from Separius/patch-1

  Show description for 368211d
  rwightman

  authored
  368211d

  Copy full SHA for 368211d

Commits on Aug 12, 2021

• Add `mmax` config key to auto_augment for increasing upper bound of RandAugment magnitude beyond 10. Make AugMix uniform sampling default not override config setting.

  rwightman

  committed
  3cdaf5e

  Copy full SHA for 3cdaf5e

Commits on Aug 9, 2021

• Add non fused LAMB optimizer option

  rwightman

  committed
  1042b8a

  Copy full SHA for 1042b8a

Commits on Aug 7, 2021

• Add gc_efficientnetv2_rw_t weights (global context instead of SE attn). Add TF XL weights even though the fine-tuned ones don't validate that well. Change default arg for GlobalContext to use scal …

  Show description for 01cb46a
  rwightman

  committed
  01cb46a

  Copy full SHA for 01cb46a

Commits on Jul 28, 2021

• Update README.md

  rwightman

  authored
  bd56946

  Copy full SHA for bd56946

Commits on Jul 22, 2021

• Add EfficientNetV2 XL model defs

  rwightman

  committed
  d3f7440

  Copy full SHA for d3f7440

Commits on Jul 13, 2021

• Attempt to fix xcit test failures on github runner by filter largest models

  rwightman

  committed
  ef1e2e1

  Copy full SHA for ef1e2e1

• Merge pull request #750 from drjinying/master

  Show description for 72b227d
  rwightman

  authored
  72b227d

  Copy full SHA for 72b227d

• Merge pull request #746 from samarth4149/master

  Show description for 2907c1f
  rwightman

  authored
  2907c1f

  Copy full SHA for 2907c1f

Commits on Jul 12, 2021

• Update README.md

  rwightman

  committed
  5aca7c0

  Copy full SHA for 5aca7c0

• Merge branch 'alexander-soare-xcit'

  rwightman

  committed
  763329f

  Copy full SHA for 763329f

• Allow act_layer switch for xcit, fix in_chans for some variants

  rwightman

  committed
  748ab85

  Copy full SHA for 748ab85

• Merge branch 'xcit' of https://github.com/alexander-soare/pytorch-image-models into alexander-soare-xcit

  rwightman

  committed
  d3255ad

  Copy full SHA for d3255ad

Commits on Jul 11, 2021

• Fix gc effv2 model cfg name

  rwightman

  committed
  f8039c7

  Copy full SHA for f8039c7

Commits on Jul 9, 2021

• Add efficientnetv2_rw_t defs w/ weights, and gc variant, as well as gcresnet26ts for experiments. Version 0.4.13

  rwightman

  committed
  392368e

  Copy full SHA for 392368e

Commits on Jul 8, 2021

• Add SAM pretrained model defs/weights for ViT B16 and B32 models.

  rwightman

  committed
  6d8272e

  Copy full SHA for 6d8272e

Commits on Jul 6, 2021

• Remove unecessary line from nest post refactor

  rwightman

  committed
  ee4d8fc

  Copy full SHA for ee4d8fc

• Merge branch 'alexander-soare-nested_transformer'

  rwightman

  committed
  c8ec1ff

  Copy full SHA for c8ec1ff

• Realized LayerNorm2d won't work in all cases as is, fixed.

  rwightman

  committed
  8165cac

  Copy full SHA for 8165cac

• Move aggregation (convpool) for nest into NestLevel, cleanup and enable features_only use. Finalize weight url.

  rwightman

  committed
  81cd686

  Copy full SHA for 81cd686

Commits on Jul 3, 2021

• Merge branch 'nested_transformer' of https://github.com/alexander-soare/pytorch-image-models into alexander-soare-nested_transformer

  rwightman

  committed
  6ae0ac6

  Copy full SHA for 6ae0ac6

Pagination

Previous

Next