Tag
Introduces N_0-VTLA, a vision-tactile-language-action foundation model for contact-rich manipulation, featuring large-scale tactile pretraining and advantage-conditioned offline policy improvement, with strong results on real-robot and simulation benchmarks.