Commits

Branch selector

 74a04e0

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 19, 2022

• Add parameter to change normalization type

  ChristophReich1996

  committed
  74a04e0

  Copy full SHA for 74a04e0

• Create model functions

  ChristophReich1996

  committed
  2a4f6c1

  Copy full SHA for 2a4f6c1

• Add get and reset classifier method

  ChristophReich1996

  committed
  87b4d7a

  Copy full SHA for 87b4d7a

• Check input resolution

  ChristophReich1996

  committed
  ff5f6bc

  Copy full SHA for ff5f6bc

• Change parameter names to match Swin V1

  ChristophReich1996

  committed
  81bf0b4

  Copy full SHA for 81bf0b4

• Add initials (CR) to model and file

  ChristophReich1996

  committed
  f227b88

  Copy full SHA for f227b88

• Add code from https://github.com/ChristophReich1996/Swin-Transformer-V2 and change docstring style to match timm

  ChristophReich1996

  committed
  90dc74c

  Copy full SHA for 90dc74c

Commits on Feb 2, 2022

• Update README.md

  rwightman

  authored
  7c67d6a

  Copy full SHA for 7c67d6a

Commits on Jan 24, 2022

• Add vit_base2_patch32_256 for a model between base_patch16 and patch32 with a slightly larger img size and width

  rwightman

  committed
  07379c6

  Copy full SHA for 07379c6

• Update benchmark and validate scripts to output results in JSON with a fixed delimiter for use in multi-process launcher

  rwightman

  committed
  cf43343

  Copy full SHA for cf43343

Commits on Jan 23, 2022

• Add train benchmark results, adjust name scheme for inference and train benchmark files.

  rwightman

  committed
  1331c14

  Copy full SHA for 1331c14

Commits on Jan 21, 2022

• Merge pull request huggingface#1105 from kozistr/refactor/remove-condition

  Show description for a517bf6
  rwightman

  authored
  a517bf6

  Copy full SHA for a517bf6

• refactor: remove if-condition

  kozistr

  committed
  56a6b38

  Copy full SHA for 56a6b38

• version 0.5.5

  rwightman

  committed
  4476776

  Copy full SHA for 4476776

Commits on Jan 20, 2022

• Add update NCHW and NHWC inference benchmark numbers for current models. Flip name of 'sam' vit models in results files

  rwightman

  committed
  499c474

  Copy full SHA for 499c474

Commits on Jan 19, 2022

• Last batch of small model weights (for now). mobilenetv3_small 050/075/100 and updated mnasnet_small with lambc/lamb optimizer.

  rwightman

  committed
  83b40c5

  Copy full SHA for 83b40c5

• Merge pull request huggingface#1094 from Mi-Peng/lars

  Show description for 7f73252
  rwightman

  authored
  7f73252

  Copy full SHA for 7f73252

• fix lars

  

  Mi-Peng

  committed
  cdcd0a9

  Copy full SHA for cdcd0a9

• Update results csvs for latest release

  rwightman

  committed
  2d4b7e7

  Copy full SHA for 2d4b7e7

• Add AvgPool2d anti-aliasing support to ResNet arch (as per OpenAI CLIP models), add a few blur aa models as well

  rwightman

  committed
  1aa617c

  Copy full SHA for 1aa617c

Commits on Jan 17, 2022

• Add --fuser arg to train/validate/benchmark scripts to select jit fuser type

  rwightman

  committed
  f0f9ecc

  Copy full SHA for f0f9ecc

• Add Dino pretrained weights (no head) for vit models. Add support to tests and helpers for models w/ no classifier (num_classes=0 in pretrained cfg)

  rwightman

  committed
  010b486

  Copy full SHA for 010b486

• unbiased=False for torch.var_mean path of ConvNeXt LN. Fix huggingface#1090

  rwightman

  committed
  738a9cd

  Copy full SHA for 738a9cd

Commits on Jan 16, 2022

• Default conv_mlp to False across the board for ConvNeXt, causing issues on more setups than it's improving right now...

  rwightman

  committed
  e0c4eec

  Copy full SHA for e0c4eec

Commits on Jan 15, 2022

• Add ConvNeXt 22k->1k fine-tuned and 384 22k-1k fine-tuned weights after testing

  rwightman

  committed
  b669f4a

  Copy full SHA for b669f4a

• Update README.md

  rwightman

  authored
  6dcbaf2

  Copy full SHA for 6dcbaf2

• Giant/gigantic vits snuck through in a test a broke GitHub test runner, add filter

  rwightman

  committed
  a8d103e

  Copy full SHA for a8d103e

• Extra vit_huge model likely to cause test issue (non in21k variant), adding to filters

  rwightman

  committed
  ef72ad4

  Copy full SHA for ef72ad4

• Update REAMDE.md. Sneak in g/G (giant / gigantic?) ViT defs from scaling paper

  rwightman

  committed
  e967c72

  Copy full SHA for e967c72

• Add some more small model weights lcnet, mnas, mnv2

  rwightman

  committed
  9ca3437

  Copy full SHA for 9ca3437

• Version 0.5.4

  rwightman

  committed
  fa6463c

  Copy full SHA for fa6463c

• Fix stem width for really small mobilenetv3 arch defs

  rwightman

  committed
  fa81164

  Copy full SHA for fa81164

• Add missing dropout for head reset in ConvNeXt default head

  rwightman

  committed
  edd3d73

  Copy full SHA for edd3d73

• Some convnext cleanup, remove in place mul_ for gamma, breaking symbolic trace, cleanup head a bit...

  rwightman

  committed
  b093dcb

  Copy full SHA for b093dcb

• Add initial ConvNeXt impl (mods of official code)

  rwightman

  committed
  18934de

  Copy full SHA for 18934de

Pagination

Previous

Next