Learning Attention Propagation for Compositional Zero-Shot Learning

This is the un-official PyTorch code of the WACV2023 paper Learning Attention Propagation for Compositional Zero-Shot Learning. The code provides the implementation of the Cape method.

Setup

1. Clone the repo

2. We recommend using Anaconda for environment setup. To create the environment and activate it, please run:

    conda env create --file environment.yml
    conda activate czsl

4. Go to the cloned repo and open a terminal. Download the datasets and embeddings, specifying the desired path (e.g. DATA_ROOT in the example):

    bash ./utils/download_data.sh DATA_ROOT
    mkdir logs

Training

    python train.py --config CONFIG_FILE

where CONFIG_FILE is the path to the configuration file of the model. The folder configs contains configuration files for all methods.

To run Cape on MitStates, the command is just:

    python train.py --config configs/cape/mit.yml

On UT-Zappos, the command is:

    python train.py --config configs/cape/utzappos.yml

On CGQA, the command is:

    python train.py --config configs/cape/cgqa.yml

Note: To create a new config, all the available arguments are indicated in flags.py.

Test

To test a model, the code is simple:

    python test.py --logpath LOG_DIR

where LOG_DIR is the directory containing the logs of a model. For example, mitstates, utzappos, cgqa.

References

@inproceedings{khan2023learning,
  title={Learning Attention Propagation for Compositional Zero-Shot Learning},
  author={Khan, Muhammad Gul Zain Ali and Naeem, Muhammad Ferjad and Van Gool, Luc and Pagani, Alain and Stricker, Didier and Afzal, Muhammad Zeshan},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages={3828--3837},
  year={2023}
}

Note: For this code I have taken halp from following github repo: https://github.com/ExplainableML/czsl