mbazaNLP/kinyarwanda-coqui-stt-model

Model card - Kinyarwanda coqui STT model

Model details

• Kinyarwanda Speech to text model
• Developed by Digital Umuganda
• Model based from: Baidu Deepspeech end to end RNN model
• paper: deepspeech end to end STT
• Documentation on model: deepspeech documentation
• License: Mozilla 2.0 License
• Feedback on the model: [email protected]

Intended use cases

• Intended to be used for
  • simple keyword spotting
  • simple transcribing
  • transfer learning for better kinyarwanda and african language models
• Intended to be used by:
  • App developpers
  • various organizations who want to transcribe kinyarwanda recordings
  • ML researchers
  • other researchers in Kinyarwanda and tech usage in kinyarwanda (e.g. Linguists, journalists)
• Not intended to be used as:
  • a fully fledged voice assistant
  • voice recognition application
  • Multiple languages STT
  • language detection

Factors

• Anti-bias: these are bias that can influence the accuracy of the model
  • Gender
  • accents and dialects
  • age
• Voice quality: factors that can influence the accuracy of the model
  • Background noise
  • short sentences
• Voice format: voices must be converted to the wav format
  • wav format

Metrics

• word error rate on the Common Voice Kinyarwanda test set

Test Corpus	WER
Common Voice	39.1%

Training data

• common voice crowdsource website

Evaluation data

• common voice crowdsource website