Search papers, labs, and topics across Lattice.
4
0
5
4
ReToken boosts vision-language model performance by over 20% on key benchmarks while fitting training and inference within a single GPU's memory constraints.
Achieving 19 times lower inference latency than diffusion models, UniGen-AR redefines the efficiency of unified visual generation.
Kimi K3's innovative architecture achieves a 2.5x scaling efficiency improvement, enabling robust performance across diverse long-horizon tasks.
AlphaWiSE achieves superior continual adaptation in multimodal models without sacrificing cross-modal alignment, outperforming traditional methods.
