metadata

language:
  - sr
tags:
  - audio
  - automatic-speech-recognition
license: mit
library_name: ctranslate2
datasets:
  - google/fleurs
  - mozilla-foundation/common_voice_16_1
  - Sagicc/audio-lmb-ds
  - espnet/yodas

Whisper Medium Yodas

This model is a fine-tuned version of openai/whisper-medium on the multiple datasets.

Testing new dataset espnet/yodas

It achieves the following results on the evaluation set:

Loss: 0.1466
Wer Ortho: 0.1771
Wer: 0.0811