Whisper Small FIFA_commentary
This model is a fine-tuned version of openai/whisper-small on the FIFA_commentary dataset. It achieves the following results on the evaluation set:
- Loss: 1.3021
- Wer: 23.0032
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: 5e-06
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 1.0 | 3 | 1.3065 | 23.6422 |
| No log | 2.0 | 6 | 1.3063 | 23.6422 |
| No log | 3.0 | 9 | 1.3060 | 23.6422 |
| No log | 4.0 | 12 | 1.3054 | 23.6422 |
| No log | 5.0 | 15 | 1.3047 | 23.6422 |
| No log | 6.0 | 18 | 1.3041 | 23.6422 |
| No log | 7.0 | 21 | 1.3038 | 23.3227 |
| No log | 8.0 | 24 | 1.3030 | 23.0032 |
| No log | 9.0 | 27 | 1.3025 | 23.0032 |
| No log | 10.0 | 30 | 1.3021 | 23.0032 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 0