XanderJC / scalable-birl

Scalable Bayesian Inverse Reinforcement Learning (ICLR 2021) by Alex J. Chan and Mihaela van der Schaar.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

td error

wadaniel opened this issue · comments

Hi there, I guess you missed the discount factor while computing the TD error:
https://github.com/XanderJC/scalable-birl/blob/main/sbirl/models.py#L189

I found that you are not using a discount factor, respectively gamma equals 1.0