Skip to content

Cleanup, XLA model compat for ResNetV2-BiT / NFNet, ECA-NFNet-L2 #702

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 17 commits into from
Jun 21, 2021
Merged

Cleanup, XLA model compat for ResNetV2-BiT / NFNet, ECA-NFNet-L2 #702

Changes from 1 commit
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
8e4ac35
All ScaledStdConv and StdConv uses default to using F.layernorm so th…
rwightman Jun 8, 2021
2f5ed2d
Update `init_values` const for 24 and 36 layer ResMLP models
rwightman Jun 8, 2021
ba2ca4b
One codepath for stdconv, switch layernorm to batchnorm so gain inclu…
rwightman Jun 12, 2021
8880f69
Refactoring, cleanup, improved test coverage.
rwightman Jun 12, 2021
4d96165
Merge branch 'master' into cleanup_xla_model_fixes
rwightman Jun 13, 2021
0020268
Try lower max size for non_std default_cfg test
rwightman Jun 13, 2021
8319e0c
Add file docstring to std_conv.py
rwightman Jun 13, 2021
b9cfb64
Support npz custom load for vision transformer hybrid models. Add pos…
rwightman Jun 14, 2021
511a8e8
Add official ResMLP weights.
rwightman Jun 15, 2021
1228f5a
Add BiT distilled 50x1 and teacher 152x2 models from 'A good teacher …
rwightman Jun 18, 2021
8257b86
Fix up resnetv2 bit/bitm model default res
rwightman Jun 18, 2021
b319eb5
Update ViT weights, more details to be added before merge.
rwightman Jun 18, 2021
4c09a2f
Bump version 0.4.12
rwightman Jun 18, 2021
8f4a022
Add GMixer-24 MLP model weights, trained w/ TPU + PyTorch XLA
rwightman Jun 18, 2021
26f04a8
Fix a weight link
rwightman Jun 19, 2021
381b279
Add hybrid model fwds back
rwightman Jun 20, 2021
9c9755a
AugReg release
rwightman Jun 21, 2021
File filter

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Fix a weight link
  • Loading branch information
@rwightman
rwightman committed Jun 19, 2021
commit 26f04a8e3ef7c581f4766cd34e71d94105c32064
2 changes: 1 addition & 1 deletion timm/models/vision_transformer_hybrid.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,7 @@ def _cfg(url='', **kwargs):

# hybrid in-21k models (weights from official Google JAX impl where they exist)
'vit_tiny_r_s16_p8_224_in21k': _cfg(
url='https://storage.googleapis.com/vit_models/augreg/Ti_16-i1k-300ep-lr_0.001-aug_medium1-wd_0.1-do_0.0-sd_0.0.npz',
url='https://storage.googleapis.com/vit_models/augreg/R_Ti_16-i21k-300ep-lr_0.001-aug_none-wd_0.03-do_0.0-sd_0.0.npz',
num_classes=21843, crop_pct=0.9, first_conv='patch_embed.backbone.conv'),
'vit_small_r26_s32_224_in21k': _cfg(
url='https://storage.googleapis.com/vit_models/augreg/R26_S_32-i21k-300ep-lr_0.001-aug_medium2-wd_0.03-do_0.0-sd_0.0.npz',
Expand Down