<!-- 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-tiny-v2-ta_e6
This model is a fine-tuned version of openai/whisper-tiny on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5092
- Wer: 56.9007
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: 1.5e-06
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.1589 | 0.11 | 300 | 1.1044 | 244.4552 |
| 0.8824 | 0.21 | 600 | 0.8741 | 111.4770 |
| 0.7377 | 0.32 | 900 | 0.6424 | 59.2252 |
| 0.6134 | 0.42 | 1200 | 0.5511 | 55.1090 |
| 0.5779 | 0.53 | 1500 | 0.5312 | 60.0969 |
| 0.558 | 0.63 | 1800 | 0.5183 | 58.0145 |
| 0.5654 | 0.74 | 2100 | 0.5092 | 56.9007 |
Framework versions
- Transformers 4.28.0.dev0
- Pytorch 1.12.1
- Datasets 2.11.0
- Tokenizers 0.13.3
