Auto Tagging for Fashion Retail Using Multi-label Image Classification

---

How to Run This App

Via Cloning This Repo

• Clone this repo: https://github.com/danielsyahputra/pytorch-image-tagging.git
• Change the working directory: cd pytorch-image-tagging
• Install dependencies: pip install -r requirements.txt
• Run thi command: streamlit run App.py

Via Docker

The image is avaliable at Docker hub. To run this app, you can do the following commands.

• Pull image: docker pull danielsyahputra13/pytorch_image_tagging
• Run docker container locally: docker run -it -rm -p {LOCAL_PORT}:8501 danielsyahputra13/streamlit_word_segmentation
• Open the browser and go to localhost:{LOCAL_PORT} to see the application.

Once all the process has successfully done, you will see the app look like this.

Citation

@inproceedings{lefakis2018feidegger,
  title={FEIDEGGER: A Multi-modal Corpus of Fashion Images and Descriptions in German},
  author={Lefakis, Leonidas and Akbik, Alan and Vollgraf, Roland},
  booktitle = {{LREC} 2018, 11th Language Resources and Evaluation Conference},
  year      = {2018}
}