Automatic Speech Recognition on GigaSpeech (dev)
0.0992WERCR-CTC/AED, Zipformer-XL
Evaluation Results
| Method | Links | |
|---|---|---|
| CR-CTC/AED, Zipformer-XLParams (M)=315.52024.10 | 0.0992 | |
| Pruned transducer w/ CR-CTC, Zipformer-XLParams (M)=286.62024.10 | 0.0995 | |
| Pruned transducer, Zipformer-XLParams (M)=288.22024.10 | 0.1009 | |
| CR-CTC, Zipformer-XLParams (M)=286.62024.10 | 0.1015 | |
| CTC/AED, Zipformer-XLParams (M)=315.52024.10 | 0.1022 | |
| Pruned transducer, Zipformer-LParams (M)=148.42024.10 | 0.1023 | |
| CTC/AED, Zipformer-LParams (M)=174.32024.10 | 0.1026 | |
| CR-CTC, Zipformer-LParams (M)=1472024.10 | 0.1031 | |
| Pruned transducer, Zipformer-MParams (M)=65.62024.10 | 0.1037 | |
| CTC/AED, Zipformer-MParams (M)=902024.10 | 0.1057 | |
| CTC/AED, E-Branchformer in ESPnetParams (M)=148.9, Reference=Watanabe et al., 20182024.10 | 0.106 | |
| CR-CTC, Zipformer-MParams (M)=64.32024.10 | 0.1062 | |
| CTC/AED, Conformer in WenetParams (M)=113.2, Reference=Zhang et al., 20222024.10 | 0.107 | |
| CTC, Zipformer-XLParams (M)=286.62024.10 | 0.108 | |
| CTC/AED, Conformer in ESPnetParams (M)=113.2, Reference=Chen et al., 2021a2024.10 | 0.109 | |
| Pruned transducer, Zipformer-SParams (M)=23.32024.10 | 0.1098 | |
| CTC, Zipformer-LParams (M)=1472024.10 | 0.1116 | |
| CTC, Zipformer-MParams (M)=64.32024.10 | 0.1123 | |
| CTC/AED, Zipformer-SParams (M)=46.32024.10 | 0.114 | |
| CR-CTC, Zipformer-SParams (M)=22.12024.10 | 0.1168 | |
| CTC, Zipformer-SParams (M)=22.12024.10 | 0.1208 | |
| CTC/AED, TransformerParams (M)=87, Reference=Chen et al., 2021a2024.10 | 0.123 |