Wav2Vec2 Large XLSR 53 Gujarati Fine tuned facebook/wav2vec2 large xlsr 53 on Gujarati using the OpenSLR SLR78 dataset. When using this model, make sure that your speech input is sampled at 16kHz. Usage The model can be used directly (without a language model) as follows, assuming you have a dataset with Gujarati sentence and path fields: Evaluation The model can be evaluated as follows on 10% of the Marathi data on OpenSLR. Test Result : 23.55 % Training 90% of the OpenSLR Gujarati Male+Female dataset was used for training, after removing few examples that contained Roman characters. The colab notebook used for training can be found here.
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy