| [Latent Memory Palace](https://github.com/WEIRDLabUW/latent-memory-palace) instantiates iterative and adaptive reasoning for control policies through variational inference with an autoregressive latent distribution. This repository contains the DROID multitask policy checkpoint. |