Vosk Tajik Speech Recognition Model
Tajik automatic speech recognition (ASR) model for sherpa-onnx. Architecture is non-streaming Zipformer2.
Version 0.61
Word Error Rate (WER)
| Dataset | WER |
|---|---|
| FLEURS Tajik | 13.85 |
| Peacockery/tajik-asr-corpus-v3 (test split) | 28.21 |
Training Data
This model was trained on Peacockery/tajik-asr-corpus-v3.
Dataset used to train alphacep/vosk-model-tg
Evaluation results
- Test WER on FLEURS Tajiktest set self-reported13.850
- Test WER on Peacockery/tajik-asr-corpus-v3test set self-reported28.210