Spoken Language Identification on VoxForge
96.3Accuracy2D ConvNet
Evaluation results
| Method | Links | |
|---|---|---|
| 2D ConvNetFeatures=log-Mel, Languages=En, Fr, De, Es, Year=20192019.10 | 96.3 | |
| 2D ConvNetFeatures=log-Mel, Languages=En, Fr, De, Es, Ru, It, Year=20192019.10 | 95.4 | |
| 2D ConvNet-Bi-directional GRU-AttentionFeatures=log-Mel, Languages=En, Fr, De, Es, Ru, It, Year=20192019.10 | 95 | |
| 1D ConvNetFeatures=Raw Audio, Languages=En, Fr, De, Es, Ru, It, Year=20192019.10 | 93.7 | |
| CNN-TDNNFeatures=log-Mel, Languages=En, Fr, De, Year=20092019.10 | 91.2 | |
| ResNet50Features=log-Mel, Languages=En, Fr, De, Es, Ru, It, Year=20192019.10 | 89 | |
| SVM-HMM modelFeatures=Not defined, Languages=En, Fr, Es, De, Year=20182019.10 | 70 |