# Low-resource speech recognition

An automatic speech recognition model based on the Whisper architecture, fine-tuned for specific languages or dialects

KinyaWhisper is a fine-tuned Kinyarwanda automatic speech recognition (ASR) system based on OpenAI's Whisper model, specifically designed for low-resource indigenous languages.

This model is a fine-tuned version of facebook/mms-1b-all on the Isma/alffa_wolof dataset, specifically designed for Wolof automatic speech recognition (ASR) tasks.

A speech recognition model fine-tuned based on facebook/wav2vec2-large-xlsr-53, supporting Bemba language

This model is a large-scale speech recognition model based on the Wav2Vec2 architecture, pre-trained and fine-tuned on 10 minutes of data from Libri-Light and Librispeech, using self-training objectives, suitable for 16kHz sampled speech audio.

This is an automatic speech recognition model fine-tuned from the facebook/wav2vec2-large-xlsr-53 model, specifically designed for recognizing the Sursilvan dialect of Romansh.

A fine-tuned speech recognition model for the Romansh Vallader dialect based on facebook/wav2vec2-large-xlsr-53, achieving a word error rate of 32.89%

Automatic speech recognition model fine-tuned on Tamil dataset based on facebook/wav2vec2-xls-r-300m

This is an automatic speech recognition model fine-tuned on the Common Voice 8.0 Hindi dataset based on the facebook/wav2vec2-xls-r-1b model, supporting Hindi speech-to-text tasks.

An automatic speech recognition model fine-tuned on the Common Voice 8 dataset based on facebook/wav2vec2-xls-r-300m, supporting Basque (bas).

UniSpeech is a unified speech representation learning model that combines supervised phoneme CTC learning and self-supervised learning, specifically fine-tuned for Italian.

Wav2Vec2 Base is an automatic speech recognition model pre-trained and fine-tuned on 16kHz sampled LibriSpeech audio for 100 hours.

An automatic speech recognition model for Egyptian Arabic, fine-tuned from the facebook/wav2vec2-large-xlsr-53 model, supporting speech input at 16kHz sampling rate.

This is an automatic speech recognition model fine-tuned on the Common Voice 8.0 Urdu dataset based on the XLS-R architecture, with a word error rate (WER) of 47.38.

This model is an Urdu automatic speech recognition (ASR) model fine-tuned on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - UR dataset based on facebook/wav2vec2-xls-r-300m

This model is an automatic speech recognition model fine-tuned on the Common Voice 8.0 Urdu dataset based on facebook/wav2vec2-large-xlsr-53

This model is an automatic speech recognition model fine-tuned on the Common Voice Urdu dataset based on facebook/wav2vec2-large-xlsr-53