You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* NOTE: official MaxVit weights (in1k) have been released at https://github.com/google-research/maxvit -- some extra work is needed to port and adapt since my impl was created independently of theirs and has a few small differences + the whole TF same padding fun.
249
-
249
+
250
250
### Sept 23, 2022
251
251
* LAION-2B CLIP image towers supported as pretrained backbones for fine-tune or features (no classifier)
252
252
* vit_base_patch32_224_clip_laion2b
@@ -277,7 +277,7 @@ And a big thanks to all GitHub sponsors who helped with some of my costs before
* (T) = TPU trained with `bits_and_tpu` branch training code, (G) = GPU trained
282
282
* GCVit (weights adapted from https://github.com/NVlabs/GCVit, code 100% `timm` re-write for license purposes)
283
283
* MViT-V2 (multi-scale vit, adapted from https://github.com/facebookresearch/mvit)
@@ -292,7 +292,7 @@ And a big thanks to all GitHub sponsors who helped with some of my costs before
292
292
*`convnext_atto_ols` - 75.9 @ 224, 77.2 @ 288
293
293
294
294
### Aug 5, 2022
295
-
* More custom ConvNeXt smaller model defs with weights
295
+
* More custom ConvNeXt smaller model defs with weights
296
296
*`convnext_femto` - 77.5 @ 224, 78.7 @ 288
297
297
*`convnext_femto_ols` - 77.9 @ 224, 78.9 @ 288
298
298
*`convnext_pico` - 79.5 @ 224, 80.4 @ 288
@@ -313,7 +313,7 @@ And a big thanks to all GitHub sponsors who helped with some of my costs before
313
313
*`cs3sedarknet_x` - 82.2 @ 256, 82.7 @ 288
314
314
*`cs3edgenet_x` - 82.2 @ 256, 82.7 @ 288
315
315
*`cs3se_edgenet_x` - 82.8 @ 256, 83.5 @ 320
316
-
*`cs3*` weights above all trained on TPU w/ `bits_and_tpu` branch. Thanks to TRC program!
316
+
*`cs3*` weights above all trained on TPU w/ `bits_and_tpu` branch. Thanks to TRC program!
317
317
* Add output_stride=8 and 16 support to ConvNeXt (dilation)
318
318
* deit3 models not being able to resize pos_emb fixed
319
319
* Version 0.6.7 PyPi release (/w above bug fixes and new weighs since 0.6.5)
@@ -346,8 +346,8 @@ More models, more fixes
346
346
* Hugging Face Hub support fixes verified, demo notebook TBA
347
347
* Pretrained weights / configs can be loaded externally (ie from local disk) w/ support for head adaptation.
348
348
* Add support to change image extensions scanned by `timm` datasets/readers. See (https://github.com/rwightman/pytorch-image-models/pull/1274#issuecomment-1178303103)
349
-
* Default ConvNeXt LayerNorm impl to use `F.layer_norm(x.permute(0, 2, 3, 1), ...).permute(0, 3, 1, 2)` via `LayerNorm2d` in all cases.
350
-
* a bit slower than previous custom impl on some hardware (ie Ampere w/ CL), but overall fewer regressions across wider HW / PyTorch version ranges.
349
+
* Default ConvNeXt LayerNorm impl to use `F.layer_norm(x.permute(0, 2, 3, 1), ...).permute(0, 3, 1, 2)` via `LayerNorm2d` in all cases.
350
+
* a bit slower than previous custom impl on some hardware (ie Ampere w/ CL), but overall fewer regressions across wider HW / PyTorch version ranges.
351
351
* previous impl exists as `LayerNormExp2d` in `models/layers/norm.py`
352
352
* Numerous bug fixes
353
353
* Currently testing for imminent PyPi 0.6.x release
@@ -580,7 +580,7 @@ Several (less common) features that I often utilize in my projects are included.
580
580
* AutoAugment (https://arxiv.org/abs/1805.09501) and RandAugment (https://arxiv.org/abs/1909.13719) ImageNet configurations modeled after impl for EfficientNet training (https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/autoaugment.py)
581
581
* AugMix w/ JSD loss (https://arxiv.org/abs/1912.02781), JSD w/ clean + augmented mixing support works with AutoAugment and RandAugment as well
582
582
* SplitBachNorm - allows splitting batch norm layers between clean and augmented (auxiliary batch norm) data
* Space-to-Depth by [mrT23](https://github.com/mrT23/TResNet/blob/master/src/models/tresnet/layers/space_to_depth.py) (https://arxiv.org/abs/1801.04590) -- original paper?
@@ -617,7 +617,7 @@ The official documentation can be found at https://huggingface.co/docs/hub/timm.
617
617
618
618
## Train, Validation, Inference Scripts
619
619
620
-
The root folder of the repository contains reference train, validation, and inference scripts that work with the included models and other features of this repository. They are adaptable for other datasets and use cases with a little hacking. See [documentation(https://huggingface.co/docs/timm/training_script).
620
+
The root folder of the repository contains reference train, validation, and inference scripts that work with the included models and other features of this repository. They are adaptable for other datasets and use cases with a little hacking. See [documentation](https://huggingface.co/docs/timm/training_script).
0 commit comments