PAI-BPR

State of the art fashion recommendation system capturing user preference and capability of attribute preference interpretation

• The whole project was implemented on Google Colab, hence it is suggested to run it there itself!!

folder structure for Code files

• Attribute_keras.ipynb - Attribute Representation Model Training
• GetAttribute.ipynb - Getting Visual representations from the Attribute Model
• GPBPR2.py - The Model File.
• train.py - Training file for our state of the art model.
• test.py - Testing our model
• Testing_Model.pynb - Notebook visualising our SOTA model and its predictions and outputs
• main.ipynb - To run train.py with the desired requirements

• “Attribute_keras.ipynb” and “GetAttribute.ipynb” extract visual features from an image .
• For the dataset please contact the original curators from this Paper

Use Citation

@misc{sagar2020paibpr,
      title={PAI-BPR: Personalized Outfit Recommendation Scheme with Attribute-wise Interpretability}, 
      author={Dikshant Sagar and Jatin Garg and Prarthana Kansal and Sejal Bhalla and Rajiv Ratn Shah and Yi Yu},
      year={2020},
      eprint={2008.01780},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}