hakuhodo-technologies / scope-rl

SCOPE-RL: A python library for offline reinforcement learning, off-policy evaluation, and selection
https://scope-rl.readthedocs.io/en/latest/
Apache License 2.0
109 stars 11 forks source link

Compatibility with Gymnasium #29

Open Joakim-Blach-Andersen opened 6 months ago

Joakim-Blach-Andersen commented 6 months ago

Scope RL doesn't appear to be compatible with Gymnasium - only with Gym which is no longer maintained.

dataset = SyntheticDataset( ... env=env, ... max_episode_steps=custom_env.step_per_episode, ... ) Traceback (most recent call last): File "", line 1, in File "", line 9, in init File "/Users/joakimblach/Documents/PhD/research-efficient-off-policy-evaluation/env/lib/python3.11/site-packages/scope_rl/dataset/synthetic.py", line 206, in __post_init__ raise ValueError( ValueError: env must be a child class of gym.Env