Multi-class Classification on IFLYTEK
78.82AccuracyNSP-BERT
Evaluation results
| Method | Links | |
|---|---|---|
| NSP-BERTBackbone=BERT2022.09 | 78.82 | |
| Resnet&Bi-LSTMParameters (MB)=6.66272026.06 | 76.58 | |
| Speech-Driven End-to-End Language Discrimination frameworkParameters (MB)=0.25232026.06 | 75.84 | |
| x-vectorParameters (MB)=4.31002026.06 | 75.76 | |
| Zero-PET2022.09 | 75.29 | |
| i-vectorParameters (MB)=2.10002026.06 | 74.3 | |
| LSTMParameters (MB)=0.08662026.06 | 74.28 | |
| PerplectionBackbone=RoBERTa, Template=Manual2022.09 | 55.88 |