Search papers, labs, and topics across Lattice.
Renmin University of China
1
0
2
Pretraining action modules with motion priors can drastically enhance VLA model performance, achieving faster convergence and better success rates in complex robot manipulation tasks.
