Implementation for our work to model fashion influence relations among major cities around the world learned from a massive set of social media images.

This code repository contains our influence-based forecaster and several other baselines used in:

Z. Al-Halah and K. Grauman. From Paris to Berlin: Discovering Fashion Style Influences Around the World. CVPR 2020.

Project page: https://www.cs.utexas.edu/~ziad/fashion_influence.html

Paper: https://arxiv.org/abs/2004.01316

If you use this code in your research, please cite the following paper:

@inproceedings{al-halah2020,
    author = {Ziad Al-Halah and Kristen Grauman},
    title = {{From Paris to Berlin: Discovering Fashion Style Influences Around the World}},
    year = {2020},
    booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    month = {June},
    doi = {10.1109/cvpr42600.2020.01015}
    arxivId = {2004.01316}
}

1. Clone this github repository.

   git clone https://github.com/ziadalh/cvpr20_city_influence.git
   cd cvpr20_city_influence

2. Install Dependencies

   conda create -n cityinfl python=3.6
   conda activate cityinfl
   pip3 install -r requirements.txt

Go to the project page and download the style trends used in this work

Run the code by pointing to one of the trend files you downloaded in the previous step and to an output directory. Example:

python main.py --f_traj trends_styles50_cities44_deseasonalized.pkl --d_output outputs

At the end of training and testing the forecasting models, you'll see the forecast errors for each model arranged together in a table similar to table 1 in the paper.