Search papers, labs, and topics across Lattice.
4
0
5
Hierarchical tokenization transforms bounding-box representation, leading to substantial improvements in localization accuracy and model performance across various benchmarks.
Achieving superior simultaneous speech translation quality without altering LLM architecture could redefine efficiency in real-time translation tasks.
PRIME-Speech achieves low-latency, accurate speech-to-speech generation without sacrificing the robust performance of existing speech-to-text models.
QueryGaussian achieves a 180x speedup in 3D instance retrieval while slashing GPU memory usage by over 70%, making it feasible on consumer-grade hardware.
