Farama-Foundation / D4RL

A collection of reference environments for offline reinforcement learning
Apache License 2.0
1.35k stars 285 forks source link

wrong trajectorys in the 'antmaze-medium-diverse-v0' dataset? #146

Closed qsa-fox closed 2 years ago

qsa-fox commented 2 years ago

when I want to get one trajectory from 'antmaze-medium-diverse-v0' dataset, it seems that some trajectorys are not terminated properly. There should be terminated in the middle of the returned trajectory, but not. That results wrong transitions. Specifically, the 75th path, indices 720 to 721, or 36035 to 36036 in total index, where the first position is (18.59402, 21.31174), and the next position become (0.80276, 0.80364), it seems to be a impossible transition.

qsa-fox commented 2 years ago

'antmaze-medium-diverse-v1' have solved this problem