marload / DistRL-TensorFlow2

🐳 Implementation of various Distributional Reinforcement Learning Algorithms using TensorFlow2.
Apache License 2.0
67 stars 9 forks source link

Huber Loss in QR-DQN #2

Open LouGalata opened 3 years ago

LouGalata commented 3 years ago

There is an error when the no_atoms is different than the batch size.

I copied Huber Loss from here And now it is working when no_atoms different than batch size.