And on top of that they say in the post that this research is going to help developers explore gameplay, but then say the model couldn’t recreate certain actions it didn’t have enough training data on

Comments