Pre-block Disfluency Prediction on DisfluencySpeech (test)
0.599Pre-block AUC616K CNN
Evaluation Results
| Method | Links | |
|---|---|---|
| 616K CNNFine-tuning=none, Input type=full utterance audio, Evaluation protocol=cross-corpus, n=4,000, Pos rate=0.6792026.04 | 0.599 |
Loading the SOTA2 catalog…