OpenWAM VTA: LIBERO-10
Predict the motion, then act. OpenWAM VTA combines video prediction and robot control in a dual-expert policy. Given the task instruction and current camera observations, it imagines how the scene should evolve and predicts the actions that realize that motion. New observations close the loop as the robot executes its plan.
This checkpoint is the step-15,000 VTA policy for LIBERO-Long (LIBERO-10). It includes both the video and action experts.
Use with OpenWAM
The release includes model weights, an inference configuration, and checksums. Use the matching configuration with the OpenWAM rollout tools. See the code repository for installation and inference, and the project page for the broader approach to video-based robot learning.
License
Model weights are released under AGPL-3.0-only. Upstream attribution and license terms are retained in NOTICE and LICENSE.upstream.