ivyyz / poar

POAR

Our paper POAR: Towards Open Vocabulary Pedestrian Attribute Recognition

train-TEMS📎

A PyTorch Lightning solution to train pedestrian open attribute recognition models

Dependencies

pip install 'git+https://github.com/katsura-jp/pytorch-cosine-annealing-with-warmup'
pip install pytorch-lightning

Usage 🚂

Finetuning 🚆

python train_finetune.py --folder data_dir --batch_size 512

Testing 🚆

CUDA_VISIBLE_DEVICES=0 python test_UDA_T2I.py --testset PA100K --trainset PETA --folder  ./test/dataset/PETA/

Training with our DataModule 📉

As long as each of the image pairs have the same stem name (i.e. img1.png and img1.txt) all that you need to do is specify the folder on runtime. Any subfolder structure will be ignored, meaning foo/bar/image1.jpg will always find its myster/folder/image1.txt so long as they share a common parent folder. All image suffixes will work, the only expectation is that captions are separated by \n.

Goal ⚽

Our aim is to create an easy to use Lightning implementation of pedestrian open attribute recognition. We will live by:

Citation

@inproceedings{Yue2023poar,
title={POAR: Towards Open Vocabulary Pedestrian Attribute Recognition},
author={Yue Zhang, Suchen Wang, Shichao Kan, Zhenyu Weng, Yigang Cen, and Yappeng Tan},
booktitle={In Proceedings of the 31st ACM International Conference on Multimedia (MM ’23), October 29–November 3, Ottawa, ON, Canada.},
year={2023},
doi={10.1145/3581783.3611719}
}

TODO ✅

• Get OpenAI's model creation script
• Create model inits
  • ResNet50
  • ResNet50x4
  • ResNet101
  • ViT-B/32
  • all models
• Create model wrapper
• Create lightning trainer
• Create dataset files
• Performance boosts
  • Mixed-precision
  • Self-distillation
  • Gradient checkpointing
  • Half-precision Adam statistics
  • Half-precision stochastically rounded text encoder weights

poar's Issues

Pre-trained checkpoints

Hello
How are you?
Thanks for contributing to this project.
Could you share the checkpoints trained on your side?
Thanks

the txt for "cls_35.txt"

Great work!
I have thie issue for the txt for "/raid2/yue/ReID/vision_language/train-CLIP-2th/train-CLIP-FT-6TScls/data/cls_35.txt".
Can you provide me for the txt, and tell me the function?
Thanks!