Awesome

Extending CLIP for Category-to-Image Retrieval in E-commerce

This repository contains the implementation and resources used for the experiments in the paper "Extending CLIP for Category-to-image Retrieval in E-commerce" published at ECIR 2022.

</div>

CLIP-ITA model configuration

Overview

This project extends the CLIP model to improve category-to-image retrieval tasks in zero-shot vs. fine-tuned settings.

Getting Started

Prerequisites

Python 3.8+
PyTorch
A GPU is recommended for training and evaluation.

Installation

Clone the repository:

git clone https://github.com/<your-repo>/clip-category-retrieval.git  
cd clip-category-retrieval

Install the dependencies:
```
pip install -r requirements.txt
```

License

This repository is licensed under the MIT License. Feel free to use, modify, and distribute the code. If you make significant modifications, please link back to this repository as a courtesy.

Citing and Authors

If you find this repository helpful, please consider citing our paper:

@inproceedings{hendriksen-2022-extending-clip,
author = {Hendriksen, Mariya and Bleeker, Maurits and Vakulenko, Svitlana and van Noord, Nanne and Kuiper, Ernst and de Rijke, Maarten},
booktitle = {ECIR 2022: 44th European Conference on Information Retrieval},
month = {April},
publisher = {Springer},
title = {Extending CLIP for Category-to-image Retrieval in E-commerce},
year = {2022}}