MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer

Benchmark Model Rank Results
text-to-speech-synthesis-on-librispeech-pc-test-cleanMaskGCT#10WER: 2.634Speaker Similarity: 0.687
text-to-speech-synthesis-on-seed-tts-test-enMaskGCT#18WER: 2.623Speaker Similarity: 0.717
text-to-speech-synthesis-on-seed-tts-test-zhMaskGCT#20CER: 2.273Speaker Similarity: 0.774