Home

Awesome

[CVPR 2024] Official Implementation of AMD

flowchart The following ranking results are produced by vit-b:

PWC PWC PWC PWC PWC PWC

Asymmetric Masked Distillation for Pre-Training Small Foundation Models<br> Zhiyu Zhao, Bingkun Huang, Sen Xing, Gangshan Wu, Yu Qiao, and Limin Wang<br> Nanjing University, Shanghai AI Lab<br>

News ๐Ÿ“ฐ

[2024.3.27] Code and models have been released!<br> [2024.2.29] Code and models will be released in the following days.<br> [2024.2.27] AMD is accpeted by CVPR2024! ๐ŸŽ‰๐ŸŽ‰๐ŸŽ‰<br>

Main Results ๐Ÿš€

โœจ Something-Something V2

MethodExtra DataBackboneResolution#Frames x Clips x CropsTop-1Top-5
AMDnoViT-S224x22416x2x370.292.5
AMDnoViT-B224x22416x2x373.394.0

โœจ Kinetics-400

MethodExtra DataBackboneResolution#Frames x Clips x CropsTop-1Top-5
AMDnoViT-S224x22416x5x380.194.5
AMDnoViT-B224x22416x5x382.295.3

โœจ AVA 2.2

MethodExtra DataExtra LabelBackbone#Frame x Sample RatemAP
AMDKinetics-400โœ—ViT-B16x429.9
AMDKinetics-400โœ“ViT-B16x433.5

โœจ UCF101 & HMDB51

MethodExtra DataBackboneUCF101HMDB51
AMDKinetics-400ViT-B97.179.6

โœจ ImageNet-1K

MethodExtra DataBackboneResolutionTop-1
AMDnoViT-S224x22482.1
AMDnoViT-B224x22484.6

Installation ๐Ÿ”จ

Please follow the instructions in INSTALL.md.

Data Preparation โžก๏ธ

Please follow the instructions in DATASET.md for data preparation.

Pre-training ๐Ÿ”„

The pre-training instruction is in PRETRAIN.md.

Fine-tuning โคด๏ธ

The fine-tuning instruction is in FINETUNE.md.

Model Zoo ๐Ÿ“

We provide pre-trained and fine-tuned models in MODEL_ZOO.md.

Acknowledgements ๐Ÿ‘

This project is built upon VideoMAEv2 and MGMAE. Thanks to the contributors of these great codebases.

Citation โœ๏ธ

If you find this repository useful, please use the following BibTeX entry for citation.

@misc{zhao2023amd,
      title={Asymmetric Masked Distillation for Pre-Training Small Foundation Models}, 
      author={Zhiyu Zhao and Bingkun Huang and Sen Xing and Gangshan Wu and Yu Qiao and Limin Wang},
      year={2023},
      eprint={2311.03149},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}