Expert¶

Description¶

Trajectories have expert data from a fine-tuned RL policy provided in the DAPG repository. The environment used to collect the dataset is AdroitHandDoor-v1.

Dataset Specs¶


Total Steps	1000000
Total Episodes	5000
Algorithm	Not provided
Author	Rodrigo de Lazcano
Email	rperezvicente@farama.org
Code Permalink	https://github.com/rodrigodelazcano/d4rl-minari-dataset-generation
Minari Version	`0.4.3` (supported)
Download	`minari download door-expert-v2`

Environment Specs¶

Note

The following table rows correspond to (in addition to the action and observation space) the Gymnasium environment specifications used to generate the dataset. To read more about what each parameter means you can have a look at the Gymnasium documentation https://gymnasium.farama.org/api/registry/#gymnasium.envs.registration.EnvSpec

This environment can be recovered from the Minari dataset as follows:

import minari

dataset = minari.load_dataset('door-expert-v2')
env  = dataset.recover_environment()


ID	AdroitHandDoor-v1
Observation Space	`Box(-inf, inf, (39,), float64)`
Action Space	`Box(-1.0, 1.0, (28,), float32)`
entry_point	`gymnasium_robotics.envs.adroit_hand.adroit_door:AdroitHandDoorEnv`
max_episode_steps	200
reward_threshold	None
nondeterministic	`False`
order_enforce	`True`
autoreset	`False`
disable_env_checker	`False`
kwargs	`{'reward_type': 'dense'}`
additional_wrappers	`()`
vector_entry_point	`None`

Evaluation Environment Specs¶

Note

This dataset doesn’t contain an eval_env_spec attribute which means that the specs of the environment used for evaluation are the same as the specs of the environment used for creating the dataset. The following calls will return the same environment:

import minari

dataset = minari.load_dataset('door-expert-v2')
env  = dataset.recover_environment()
eval_env = dataset.recover_environment(eval_env=True)

assert env.spec == eval_env.spec