This is the repo for paper: Knowledge Enhanced Neural Fashion Trend Forecasting
Please refer to the updated repo KERN_reproduce for the code and dataset, which are re-organized to be well-structured and easy to be understood.
- OS: Ubuntu 16.04 or higher version
- python3
- python modules: yaml, pytorch, tqdm, numpy, pickle
- The entry script is: train.py
- The config file is: config.yaml
- utility.py: the script for dataloader
- model: the folder for model files
-
Download the dataset, decompress it and put it on the top directory: tar -zxvf dataset.tgz Note that, the downloaded files include both proprecessed datasets of GeoStyle and FIT. If you want to download the original GeoStyle dataset and reproduce the preprocessing, kindly run the script 0_preprocess_data.py. Detailed introduction is within the script.
-
Change the hyper-parameters in the configure file config.yaml.
-
Run: train.py
This project is supported by the National Research Foundation, Prime Minister's Office, Singapore under its IRC@Singapore Funding Initiative.
