Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
2
0
5
VLMs on edge devices are surprisingly energy-intensive, with text generation dominating image processing costs, and NPU offloading drastically outperforming CPU backends.
Ditch the complex modules: a minimalist, end-to-end vision-language-action model for UAV navigation achieves 3x better generalization than leading baselines by directly mapping visual inputs and language to continuous control.