Skip to content

Commit 78a4123

Browse files
xianzhidutensorflower-gardener
authored andcommitted
Internal change
PiperOrigin-RevId: 423725861
1 parent 45848c4 commit 78a4123

File tree

2 files changed

+15
-0
lines changed

2 files changed

+15
-0
lines changed

official/README.md

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -43,6 +43,7 @@ In the near future, we will add:
4343
| [ResNet](vision/beta/MODEL_GARDEN.md) | [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) |
4444
| [ResNet-RS](vision/beta/MODEL_GARDEN.md) | [Revisiting ResNets: Improved Training and Scaling Strategies](https://arxiv.org/abs/2103.07579) |
4545
| [EfficientNet](vision/image_classification) | [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://arxiv.org/abs/1905.11946) |
46+
| [Vision Transformer](vision/beta/MODEL_GARDEN.md) | [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) |
4647

4748
#### Object Detection and Segmentation
4849

official/vision/beta/MODEL_GARDEN.md

Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -55,6 +55,20 @@ depth, label smoothing and dropout.
5555
| ResNet-RS-350 | 256x256 | 164.3 | 83.7 | 96.7 | [config](https://github.com/tensorflow/models/blob/master/official/vision/beta/configs/experiments/image_classification/imagenet_resnetrs350_i256.yaml) \| [ckpt](https://storage.cloud.google.com/tf_model_garden/vision/resnet-rs/resnet-rs-350-i256.tar.gz) |
5656
| ResNet-RS-350 | 320x320 | 164.3 | 84.2 | 96.9 | [config](https://github.com/tensorflow/models/blob/master/official/vision/beta/configs/experiments/image_classification/imagenet_resnetrs420_i256.yaml) \| [ckpt](https://storage.cloud.google.com/tf_model_garden/vision/resnet-rs/resnet-rs-350-i320.tar.gz) |
5757

58+
59+
#### Vision Transformer (ViT)
60+
61+
We support [ViT](https://arxiv.org/abs/2010.11929) and [DEIT](https://arxiv.org/abs/2012.12877) implementations in a TF
62+
Vision
63+
[project](https://github.com/tensorflow/models/tree/master/official/projects/vit). ViT models trained under the DEIT settings:
64+
65+
model | resolution | Top-1 | Top-5 |
66+
--------- | :--------: | ----: | ----: |
67+
ViT-s16 | 224x224 | 79.4 | 94.7 |
68+
ViT-b16 | 224x224 | 81.8 | 95.8 |
69+
ViT-l16 | 224x224 | 82.2 | 95.8 |
70+
71+
5872
## Object Detection and Instance Segmentation
5973

6074
### Common Settings and Notes

0 commit comments

Comments
 (0)