techiaith/whisper-large-ft-cy-en

Automatic Speech Recognition • 2B • Updated 23 days ago • 2.37k

Note Model adnabod lleferydd Cymraeg a Saesneg wedi'i fireinio ar gorpora lleferydd Cymraeg helaeth. Mae'n trawsgrifio'n gywir ar draws lleferydd digymell a darlleniadau. A Welsh and English speech recognition model, fine-tuned on extensive Welsh speech corpora. It transcribes accurately across spontaneous and read speech.

techiaith/whisper-large-ft-cy-en-ct2

Automatic Speech Recognition • Updated 23 days ago • 2.32k

Note Fersiwn CTranslate2 o'r model adnabod lleferydd Cymraeg a Saesneg, ar gyfer trawsgrifio cyflymach ac ysgafnach ar galedwedd gyfyngedig. A CTranslate2 version of the Welsh and English speech recognition model, for faster and lighter transcription on constrained hardware.

techiaith/kaldi-cy

Automatic Speech Recognition • Updated 3 days ago

Note Model adnabod lleferydd ystadegol sy'n seiliedig ar Kaldi ar gyfer y Gymraeg gyda chuddni isel iawn sy'n addas ar gyfer cymwysiadau trawsgrifio amser real. A Kaldi based, statistical speech recognition model for Welsh with very low latency suitable for real-time transcription applications.

techiaith/wav2vec2-btb-cv-ft-cv-cy

Automatic Speech Recognition • 0.3B • Updated Nov 6, 2024 • 10

Note Trawsgrifio lleferydd Cymraeg wedi'i ddarllen neu wedi'i gynllunio, gan ddefnyddio modelau acwstig sydd wedi ei seilio ar acwstig wav2vec2 ac, yn ddewisol, modelau iaith seiliedig ar KenLM. Yn darparu'r sgoriau WER isaf. Transcription of read or planned Welsh language speech using wav2vec2 acoustic based acoustic models and, optionally, KenLM based language models. Provides lowest WER scores.

techiaith/fullstop-welsh-punctuation-prediction

Token Classification • 0.6B • Updated Mar 21, 2024 • 7 • 1

Note Model sy'n gallu mewnosod atalnodi i allbwn testun adnabod lleferydd A model that can insert punctuation into speech recognition text output

techiaith/sacremoses-truecaser-cy

Updated Nov 5, 2024

Note Model sy'n gallu cywiro priflythrennu allbwn testun adnabod lleferydd A model that can correct the capitalization of speech recognition text output