Search papers, labs, and topics across Lattice.
3
0
4
2
Unified generation of temporally structured audio achieves unprecedented speaker similarity and cross-turn consistency without task-specific branches.
Achieving state-of-the-art performance in speech synthesis, Qwen-Audio-3.0-TTS excels in multilingual support and robustness against challenging audio conditions.
LSLMs can be significantly compressed without sacrificing accuracy by aggressively merging redundant tokens in deeper layers, challenging the need for fully distinct token representations.
