Atharva Mete

We compare to state-of-the-art imitation learning and L VM baselines and see that QueST's architecture leads to strong performance on several multitask and few-shot learning benchmarks.