This repository contains supplementary code for the paper "Symmetry-invariant Transformers for Generalisation in Reinforcement Learning," accepted at ICML 2024.
As our SiT model is tested on top of multiple RL backbone algorithms, we provide a step-by-step guide in the Readme files of the respective folders in this repo on how to modify the code-bases, respectively.
Under development... Currently available: the "SiT_model.py" file contains the code for respective SiT models/architecture with different symmetry properties.