Search papers, labs, and topics across Lattice.
1
0
3
By intelligently grouping vector quantized features based on their information content, EG-GRVQ achieves state-of-the-art speech reconstruction at ultra-low bitrates.
