MolmoMotion is a language-guided 3D motion forecasting model that, given a video frame, marked 3D points on an object, and written action instructions, predicts where those points will move over the next few seconds in 3D space.
Need help?
Contact usMolmoMotion is a language-guided 3D motion forecasting model that, given a video frame, marked 3D points on an object, and written action instructions, predicts where those points will move over the next few seconds in 3D space.