Whisper Base - Kiran Pantha
This model is a fine-tuned version of openai/whisper-base on the OpenSLR54 dataset. It achieves the following results on the evaluation set:
- Loss: 0.2056
- Wer: 43.2821
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: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.5029 | 0.0750 | 500 | 0.4922 | 77.0205 |
| 0.351 | 0.1499 | 1000 | 0.3561 | 65.6941 |
| 0.3034 | 0.2249 | 1500 | 0.2988 | 57.0618 |
| 0.2689 | 0.2999 | 2000 | 0.2714 | 53.2844 |
| 0.2584 | 0.3749 | 2500 | 0.2537 | 50.8369 |
| 0.2325 | 0.4498 | 3000 | 0.2393 | 48.0282 |
| 0.2238 | 0.5248 | 3500 | 0.2271 | 46.5723 |
| 0.2149 | 0.5998 | 4000 | 0.2149 | 44.4056 |
| 0.2038 | 0.6748 | 4500 | 0.2091 | 43.6834 |
| 0.2026 | 0.7497 | 5000 | 0.2056 | 43.2821 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.0+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1