[paper]
[Project Page]
[Weights]
This is a PyTorch implementation of the ICLR23 paper: LiftedCL:
@inproceedings{
chen2023liftedcl,
title={Lifted{CL}: Lifting Contrastive Learning for Human-Centric Perception},
author={Ziwei Chen and Qiang Li and Xiaofeng Wang and Wankou Yang},
booktitle={The Eleventh International Conference on Learning Representations },
year={2023},
url={https://openreview.net/forum?id=WHlt5tLz12T}
}
[04/2023] Training codes release!
[03/2023] Pre-trained ResNet-50 model (IN+CC 200 epoch) release!
[01/2023] LiftedCL has been accepted to ICLR 2023!
Pytorch (we test our codes with 1.11)
torchvision
python train_cl.py --multiprocessing-distributed ./path_to_dataset
python train_adv.py --multiprocessing-distributed ./path_to_dataset
We hope our work can inspire others when doing 3D-aware representation learning. Lifting and adv training is feasible for human-centric tasks, but there remains performance potential. Besides, how to do the 3D-aware representation learning for other tasks (e.g., Object Dection) is worth further research!
We would like to thank the MoCo for its open-source project.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent