Interpretability in realistic environments#14

Open

AdamGleave wants to merge 12 commits intomasterfrom

hidim-interpretability

Commits on Feb 3, 2020

Runner script to sparify reward models for PointMaze
AdamGleave
committed

Commits on Feb 4, 2020

Merge remote-tracking branch 'origin/master' into hidim-interpretability
AdamGleave
committed

Commits on Feb 5, 2020

Rewards: refactor to have common GymHardcodedReward base class
AdamGleave
committed
Bugfix in PointMassGroundTruth: use next observation like step()
AdamGleave
committed
Fixed LunarLander environment and ground-truth reward model
AdamGleave
committed

Commits on Mar 2, 2020

Bugfix: time limit in lunar lander
AdamGleave
committed