The original code for the paper "How to train your MAML" along with a replication of the original "Model Agnostic Meta Learning" (MAML) paper in Pytorch.
thanks for your great work. I have some doubt about the "self.args.multi_step_loss_num_epochs" in function "get_per_step_loss_importance_vector". What does it mean?
thanks for your great work. I have some doubt about the "self.args.multi_step_loss_num_epochs" in function "get_per_step_loss_importance_vector". What does it mean?