Tips is a smart GNN recommender system library which helps choose recommender systems with most optimised results.
Tips is a use-case centric library and the end-goal is to implement GNN based recommender systems with just 2-3 lines of code, compare them and protoype the best working model.
from tips import Tips, TipsFactory
properties_dict = {"model": "LightGCN", "data_path":"/path/to/data"}
new_tip = TipsFactory(properties_dict)
train, valid, test = Tips.create_dataset(new_tip)
model, best_score, valid_score = Tips.train(Tips.get_model(new_tip))
Tips.eval_model(model)An easier approach is to build directly from the properties dict.
from tips import Tips
properties_dict = {"model": "LightGCN", "data_path":"/path/to/data"}
model, best_score, valid_score = Tips.train_model(properties_dict)
Tips.eval_model(model)This library is WIP.