mozilla-ai/whisper-large-v3-turbo-bn

Finetuned openai/whisper-large-v3-turbo on 21409 Bengali training audio samples from cv-corpus-21.0-2025-03-14/bn.

This model was created from the Mozilla.ai Blueprint: speech-to-text-finetune.

Evaluation results on 9363 audio samples of Bengali:

Baseline model (before finetuning) on Bengali

• —Word Error Rate (Normalized): 78.843
• —Word Error Rate (Orthographic): 107.027
• —Character Error Rate (Normalized): 62.521
• —Character Error Rate (Orthographic): 72.012
• —Loss: 1.074

Finetuned model (after finetuning) on Bengali

• —Word Error Rate (Normalized): 11.053
• —Word Error Rate (Orthographic): 26.436
• —Character Error Rate (Normalized): 6.059
• —Character Error Rate (Orthographic): 7.537
• —Loss: 0.109