whisper-small-sr

Fine-tuned OpenAI Whisper Small.

Output script: this model is intended to produce Serbian Latin only.

• WER on Common Voice 24.0 Serbian test: 6.59%

Model description

Training and evaluation data

This model was fine-tuned on a mixture of publicly available Serbian speech corpora, including:

• Mozilla Common Voice 24.0, evaluated on CV test (sr)
• FLEURS Serbian
• ParlaSpeech-RS (subset of the full dataset)
• Additional Serbian corpora used in the training pipeline

Training procedure

• Epochs: 9
• Batch size: 32 / 20
• Optimizer: AdamW
• LR: 6e-5 with warmup (50 steps) + cosine decay to min_lr = 1e-7
• Mixed precision: bfloat16 (fp32 in the final epoch)
• SpecAugment: frequency + time masking
• Sampling: weighted sampling across datasets

Training results

Epoch	Train loss	CV WER
1	0.333	0.1614
2	0.344	0.1278
3	0.251	0.1112
4	0.202	0.1032
5	0.167	0.0934
6	0.138	0.0790
7	0.118	0.0740
8	0.103	0.0709
9	0.096	0.0659

Evaluation Metrics

• WER (normalized) on Common Voice 24.0 Serbian test: 7.09%
• Text normalization used for WER:
  • punctuation removed
  • lowercased
  • Cyrillic → Latin conversion
  • numbers converted to words