Fix weighted_mean function

tristandeleu / pytorch-maml-rl

Reinforcement Learning with Model-Agnostic Meta-Learning in Pytorch

MIT License

827 stars 158 forks source link

Closed tristandeleu closed 6 years ago

tristandeleu commented 6 years ago

Fix weighted_mean function: take the mean over weighted means in a given dimension, instead of taking the overall weighted mean.
Add dim=0 argument to weighted_mean in MetaLearner