Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
3
0
5
One-step action generation for robotic foundation models is now achievable without performance collapse, slashing inference latency by up to 74%.
CAT-Q quantizes large language models with unprecedented efficiency, achieving superior performance while slashing training token requirements by 100,000X.
Training more vision models can actually *increase* efficiency, thanks to a novel pre-training strategy that leverages knowledge transfer across a "chain" of models.