<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
whisper-large-v2-es
This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2930
- Wer: 31.3542
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
- gradient_accumulation_steps: 2
- 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: 200
- training_steps: 6000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0328 | 0.16 | 250 | 0.3149 | 26.25 |
| 0.0498 | 0.33 | 500 | 0.2925 | 37.9167 |
| 0.0666 | 0.49 | 750 | 0.2762 | 28.3333 |
| 0.0909 | 0.65 | 1000 | 0.2705 | 32.1875 |
| 0.0843 | 0.82 | 1250 | 0.2566 | 27.7083 |
| 0.1277 | 0.98 | 1500 | 0.2609 | 29.7917 |
| 0.0431 | 1.14 | 1750 | 0.2667 | 30.6250 |
| 0.0321 | 1.31 | 2000 | 0.2570 | 30.8333 |
| 0.0301 | 1.47 | 2250 | 0.2793 | 31.1458 |
| 0.0364 | 1.63 | 2500 | 0.2769 | 29.375 |
| 0.0303 | 1.8 | 2750 | 0.2755 | 28.6458 |
| 0.0226 | 1.96 | 3000 | 0.2846 | 27.9167 |
| 0.0185 | 2.12 | 3250 | 0.2863 | 26.9792 |
| 0.0161 | 2.29 | 3500 | 0.2902 | 27.8125 |
| 0.0165 | 2.45 | 3750 | 0.2903 | 32.7083 |
| 0.0145 | 2.61 | 4000 | 0.2903 | 29.8958 |
| 0.0123 | 2.78 | 4250 | 0.3076 | 27.8125 |
| 0.0129 | 2.94 | 4500 | 0.3019 | 28.4375 |
| 0.0145 | 3.1 | 4750 | 0.2917 | 26.5625 |
| 0.0118 | 3.27 | 5000 | 0.2953 | 33.6458 |
| 0.0099 | 3.43 | 5250 | 0.2881 | 33.0208 |
| 0.0088 | 3.59 | 5500 | 0.2922 | 31.5625 |
| 0.0118 | 3.76 | 5750 | 0.2935 | 30.5208 |
| 0.0103 | 3.92 | 6000 | 0.2930 | 31.3542 |
Framework versions
- Transformers 4.34.0.dev0
- Pytorch 1.12.1+cu116
- Datasets 2.4.0
- Tokenizers 0.12.1
