Search papers, labs, and topics across Lattice.
3
0
5
Hy-Embodied-VLM-1.0 outperforms its predecessor by 8.4% while activating only a fraction of the parameters, redefining efficiency in embodied agents.
ViQ achieves a groundbreaking balance between semantic richness and detail in visual representations, enabling efficient multimodal training without sacrificing quality.
Teaching VLMs to predict depth maps during pre-training unlocks surprisingly large gains in real-world robot task execution.
