Search papers, labs, and topics across Lattice.
Affiliation:
1
0
2
Achieving 2.8x compression in on-device speech tokenizers without sacrificing accuracy could revolutionize mobile AI applications.
