I want to train my own models and see I can get better then what you did, so having your datasets would be helpful