MT-Opt: Continuous Multi-Task Robotic Reinforcement Learning at Scale

Builds on top of QT-Opt.

MT-opt = multi-task optimization?

Project page:

• https://karolhausman.github.io/mt-opt/