Whisper Base Maltese
This model is a fine-tuned version of openai/whisper-base on the Common Voice 16.1 dataset. It achieves the following results on the evaluation set:
- Loss: 1.2750
- Wer: 70.1338
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: 0.001
- train_batch_size: 16
- eval_batch_size: 16
- 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: 1100
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
1.1933 | 0.9050 | 200 | 1.2337 | 81.0822 |
1.5976 | 1.8100 | 400 | 1.9209 | 93.5959 |
1.5516 | 2.7149 | 600 | 2.0278 | 91.2732 |
0.8341 | 3.6199 | 800 | 1.5194 | 82.5044 |
0.2756 | 4.5249 | 1000 | 1.2750 | 70.1338 |
Framework versions
- Transformers 4.40.0
- Pytorch 2.2.2+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1
- Downloads last month
- 7