Fine-tuned XLSR-53 large model for speech recognition in English
Fine-tuned facebook/wav2vec2-large-xlsr-53 on English using the train and validation splits of Common Voice 6.1. When using this model, make sure that your speech input is sampled at 16kHz.
- Downloads last month
- 3
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
Datasets used to train BeebekBhz/wav2vec2-large-xlsr-english
Evaluation results
- Test WER on Common Voice enself-reported19.060
- Test CER on Common Voice enself-reported7.690
- Test WER (+LM) on Common Voice enself-reported14.810
- Test CER (+LM) on Common Voice enself-reported6.840
- Dev WER on Robust Speech Event - Dev Dataself-reported27.720
- Dev CER on Robust Speech Event - Dev Dataself-reported11.650
- Dev WER (+LM) on Robust Speech Event - Dev Dataself-reported20.850
- Dev CER (+LM) on Robust Speech Event - Dev Dataself-reported11.010