We carefully design each module to make the codes friendly for newcomers. Only three steps are needed to run a model. And each module is well developed to resued, which makes building new models and creating new datasets easier.
Each model and dataset processed will be reviewed by three persons to make the code correct. All details are open sources, such as data processing, data splitting, different predictions model, and model parameters optimization.
This library is keeping updated. We will release at least one new model and one new dataset monthly before covering the most popular models and datasets. All experiments' results will be released on our website after they finish running.