ugr-sail / sinergym

Gym environment for building simulation and control using reinforcement learning
https://ugr-sail.github.io/sinergym/
MIT License
131 stars 35 forks source link

v(3.3.4) - NormalizeObservation calibration load improvement #413

Closed AlejandroCN7 closed 5 months ago

AlejandroCN7 commented 5 months ago

Description

This PR enhances the loading of previous normalization calibrations for model evaluations. It now allows specifying paths to txt files generated by Sinergym.

Additionally, this patch includes improved checks and error messages when performing these loads.

Motivation and Context

Types of changes

Checklist:

Changelog:

Extra: