D4RL icon indicating copy to clipboard operation
D4RL copied to clipboard

A collection of reference environments for offline reinforcement learning

Results 102 D4RL issues
Sort by recently updated
recently updated
newest added

Looking forward to your replay, When i tried to download the data medium-replay, I got the error: File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper File...

### Question Is there any way that can let the goal location fixed? and can set the loctaion personally?

Hi, I'm trying to generate the ant maze dataset using the generation script but getting a "No module named 'locomotion.ant'" error when loading the policy `load_policy('ant_hierarch_pol.pkl')`. I installed the locomotion...

### Question File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper File "h5py/h5f.pyx", line 106, in h5py.h5f.open OSError: Unable to open file (truncated file: eof = 5365760,...

Hi there I think there is a bug with the antmaze-umaze-diverse datasets relating to the rewards/terminal states. Versions 0 and 2. For all other datasets there is a reward/terminal of...

Hi, I want to visualize the action after training, but I notice the "d4rl/scripts/visualize_dataset.py" writed 'Only MuJoCo-based environments can be visualized', If I want to visulize 'kitchen-complete-v0',how can I do...

**Describe the bug** The behavior of `copy.deepcopy(env)` is currently undefined. I would expect either an error to be raised in case the environment should not be copied or for the...

### Question I am trying to do offline reinforcement learning with an environment that I have written myself which is derived from the Minigrid. Is it possible to generate datasets...

### Question For instance, in [ant.py:_get_obs()](https://github.com/Farama-Foundation/D4RL/blob/d842aa194b416e564e54b0730d9f934e3e32f854/d4rl/locomotion/ant.py#L90-L118) I see the observation is built from the qpos and qvel members of some physics object... but I don't understand how to qpos and...

I'm having a hard time figuring out how qlearning dataset is being built. As mentioned by @odelalleau in https://github.com/Farama-Foundation/D4RL/issues/182, the `"terminals"` key in some env is never `True`. Moreover, ```...