whisper-small-sw
This model is a fine-tuned version of smutuvi/whisper-small-sw-common-voice on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.1051
Model description
More information needed
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 1
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.0783 | 0.08 | 100 | 1.1404 |
| 1.2105 | 0.16 | 200 | 1.1282 |
| 1.189 | 0.25 | 300 | 1.1203 |
| 1.0533 | 0.33 | 400 | 1.1164 |
| 1.0932 | 0.41 | 500 | 1.1135 |
| 1.1618 | 0.49 | 600 | 1.1108 |
| 1.165 | 0.57 | 700 | 1.1090 |
| 1.1655 | 0.65 | 800 | 1.1076 |
| 1.0547 | 0.74 | 900 | 1.1066 |
| 1.1679 | 0.82 | 1000 | 1.1058 |
| 1.0539 | 0.9 | 1100 | 1.1053 |
| 1.0709 | 0.98 | 1200 | 1.1051 |
Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.1.1+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0