Search papers, labs, and topics across Lattice.
1
0
2
7
Current singing voice synthesis models struggle with genre discrimination, showing that genre-specific training can dramatically improve performance where zero-shot methods fail.
