Search papers, labs, and topics across Lattice.
1
0
2
4
Achieving a staggering reduction in Word Error Rate from 12.15% to 2.79%, MiDashengLM-Gen sets a new standard for text-to-audio generation.
