V-JEPA 2
A predictive video representation model adapted for action-conditioned robot planning.
DemonstratedCurrent
Approach
- Family
- Latent dynamics
- Weights
- Open weights
- Code
- Open code
- Paper / report
- Publicly available
- API
- Unknown
- Product
- Unknown
Evidence-backed claim
V-JEPA 2 is adapted with action conditioning for zero-shot robot planning experiments.
Source last checked:
Open primary source