Search papers, labs, and topics across Lattice.
1
0
3
Achieve state-of-the-art universal audio representation by unifying diverse audio tasks into a single next-token prediction framework, outperforming Whisper by a large margin.
