bhattasp/whisper-finetuned-atco2-uwb
This model is a fine-tuned version of openai/whisper-tiny on the AtCO2_UWB dataset. It achieves the following results on the evaluation set:
- Loss: 0.5665
- Wer: 26.4709
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: 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
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.7857 | 1.0 | 609 | 0.8737 | 51.9918 |
| 0.5851 | 2.0 | 1218 | 0.6316 | 35.6126 |
| 0.4177 | 3.0 | 1827 | 0.5783 | 29.5698 |
| 0.2927 | 4.0 | 2436 | 0.5658 | 26.8135 |
| 0.2413 | 5.0 | 3045 | 0.5665 | 26.4709 |
Framework versions
- Transformers 4.45.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0
- Downloads last month
- 77
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 bhattasp/whisper-finetuned-atco2-uwb
Base model
openai/whisper-tiny