Edit model card

🇺🇦 Join Ukrainian Speech Recognition Community - https://t.me/speech_recognition_uk

⭐ See other Ukrainian models - https://github.com/egorsmkv/speech-recognition-uk

Code for fine-tuning: https://github.com/egorsmkv/whisper-ukrainian

Tests:

  • WER: 27% (model quality is 73%)

The test with the original Whisper model is WER 30.57% (model quality is 69.43%)

Tests were with https://github.com/egorsmkv/cv10-uk-testset-clean

Downloads last month
15
Safetensors
Model size
242M params
Tensor type
F32
·

Dataset used to train Yehor/whisper-small-ukrainian

Space using Yehor/whisper-small-ukrainian 1

Evaluation results