Automatic Speech Recognition
TensorBoard
Safetensors
Welsh
wav2vec2
Generated from Trainer
DewiBrynJones's picture
Update README.md
b594d32 verified
|
raw
history blame
887 Bytes
metadata
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
  - automatic-speech-recognition
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: wav2vec2-xlsr-53-ft-btb-cv-cy-cand
    results: []
datasets:
  - techiaith/commonvoice_18_0_cy
  - techiaith/banc-trawsgrifiadau-bangor
language:
  - cy
pipeline_tag: automatic-speech-recognition

wav2vec2-xlsr-53-ft-btb-cv-cy

This model is a version of facebook/wav2vec2-large-xlsr-53 fine-tuned on Welsh language read-speech from commonvoice_18_cy and spontaneous speech from Bangor Transcriptions 24.10

It achieves the following WER results on the Bangor Transcriptions test set:

  • WER: 36.23
  • CER: 12.55