search

rdnfn / beobench

A toolkit providing easy and unified access to building control environments for reinforcement learning (RL).
https://beobench.readthedocs.io
MIT License
37 stars 4 forks source link

Enable full observation state extraction from energym integration #62

Closed rdnfn closed 2 years ago

rdnfn commented 2 years ago

Problem

Only the normalised observation state is currently accessible to the agent, and thus saveable via wandb.

Potential Solution

Pass the full observation space via the info parameter.

rdnfn commented 2 years ago

This has been implemented and will be shipped with v0.4.3 🚀

The feature can be activated by adding the following part to the config:

general:
    (...)
    log_episode_data_from_rllib: True