Search papers, labs, and topics across Lattice.
1
0
2
M* achieves up to 2.9x lower real-time factor and 2.7x higher throughput for text-to-speech tasks, revolutionizing how we serve complex multimodal AI models.
