whisper-arpa
This model is a fine-tuned version of openai/whisper-small.en on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0186
- Wer: 2.9762
- Cer: 1.0999
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-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 1
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Cer | Validation Loss | Wer |
|---|---|---|---|---|---|
| 0.0529 | 0.5716 | 5000 | 3.5532 | 0.0494 | 7.8609 |
| 0.028 | 1.1432 | 10000 | 5.5656 | 0.0413 | 9.7632 |
| 0.0298 | 1.7149 | 15000 | 2.9926 | 0.0350 | 6.4210 |
| 0.0133 | 2.2865 | 20000 | 2.1896 | 0.0322 | 5.0360 |
| 0.0105 | 2.8581 | 25000 | 2.0795 | 0.0307 | 4.7797 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 27
Inference Providers
NEW
This model is not currently available via any of the supported third-party Inference Providers, and
the model is not deployed on the HF Inference API.
Model tree for maratos/checkpoints_1
Base model
openai/whisper-small.en