-
```
env = gym.make('maze2d-umaze-v1')
dataset = d4rl.qlearning_dataset(env)
(dataset['observations'] == [0, 0, 0, 0]).sum(axis=0)
>>> array([12459, 12459, 12459, 12459])
```
What's the deal …
-
In code "cql_mujoco_new.py", define expl_path_collector
[https://github.com/aviralkumar2907/CQL/blob/master/d4rl/examples/cql_mujoco_new.py#L65](url)
[https://github.com/aviralkumar2907/CQL/blob…
-
is it possible to access the propensities (pi(a|s) for each transition s,a,r) of the logging policy while making the dataset? this would be useful for algorithms that aren't policy agnostic.
-
How to reproduce:
git clone https://github.com/rail-berkeley/d4rl.git
cd d4rl
pip install -e .
Error message text:
Exception:
Traceback (most recent call last):
File "/home/kamran/rlfd_e…
-
Greetings.
Thank you for your amazing work on Offline RL, as well as for open-sourcing the code.
This present issue pertains to the computation for the lower bounding component of the SAC CQL:
…
-
Thanks for the great work first!
I have a bunch of data in **_(state, action, reward, next state)_** format. I try to understand how you guys parse the $store$_action_ckpt file in the code but I fail…
-
I failed to run `train_brac.py` with the given `requirements.txt`. And I find the following dependencies versions work fine:
```
tensorflow==1.15.0
tensorflow-probability==0.8.0rc0
tf-agents==0.…
-
Ant maze and maze2d datasets have different numbers of samples for observations/actions/... after pulling the newest version.
@justinjfu Did you break something with your update yesterday?
-
(Crosspost of [this issue at d4rl_evaluations](https://github.com/rail-berkeley/d4rl_evaluations/issues/13))
Hi,
I find it irritating that the observations in the maze2d tasks only contain the 2…
-
Hello,
Thank you for the work on this code.
I try to train a Offline RL agent on flow, but I'm unable to save the model, I get the following error :
`Traceback (most recent call last):
Fil…