Speech-to-Text on LibriSpeech LS960 (test-other)
9.2WERMELD
Evaluation Results
| Method | Links | |
|---|---|---|
| MELDSize=260M, Hrs=960, Beam search size=52026.05 | 9.2 | |
| MELDSize=200M, Hrs=960, Beam search size=52026.05 | 10 | |
| dMel (ASR)Size=258M, Hrs=960, Beam search size=52026.05 | 10.4 | |
| MELD w/o SpecAugSize=200M, Hrs=960, Beam search size=52026.05 | 12.5 | |
| Codec-LMSize=200M, Hrs=960, Beam search size=52026.05 | 16.4 | |
| Codec-LM w/o codebook initSize=200M, Hrs=960, Beam search size=52026.05 | 100 |