Update
Use an updated fine tunned version Sagicc/whisper-medium-sr-v2 with new 10+ hours of dataset.
Whisper Medium cmb
This model is a fine-tuned version of openai/whisper-medium on the Common Voice 13 dataset. It achieves the following results on the evaluation set:
- Loss: 0.1374
- Wer Ortho: 0.1589
- Wer: 0.0658
Model description
This is a fine tunned on merged datasets Common Voice 13 + Fleurs + Juzne vesti (South news)
Rupnik, Peter and Ljubešić, Nikola, 2022,
ASR training dataset for Serbian JuzneVesti-SR v1.0, Slovenian language resource repository CLARIN.SI, ISSN 2820-4042,
http://hdl.handle.net/11356/1679.
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 1500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 0.342 | 0.48 | 500 | 0.1604 | 0.1863 | 0.0862 |
| 0.3454 | 0.95 | 1000 | 0.1388 | 0.1589 | 0.0667 |
| 0.2247 | 1.43 | 1500 | 0.1374 | 0.1589 | 0.0658 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.14.1