CALVIN

Composing Actions from Language and Vision

MIT LicenseIntroduced 2021-12-06

CALVIN (Composing Actions from Language and Vision), is an open-source simulated benchmark to learn long-horizon language-conditioned robot manipulation tasks.