N2M: Bridging Navigation and Manipulation by Learning Initial Pose Preference from Rollout

Code for Learning to Synthesize Programs as Interpretable and Generalizable Policies in NeurIPS 2021

Code and website for for SPRINT: Scalable Policy Pre-Training via Language Instruction Relabeling

Code for the paper Bootstrap Your Own Skills: Learning to Solve New Tasks with Large Language Model Guidance, accepted to CoRL 2023 as an Oral Presentation.

FurnitureBench: Real-World Furniture Assembly Benchmark (RSS 2023)

This is the official implementation of Efficient Multi-Task Learning via Selective Behavior Sharing.