Configuration Parsing
Warning:
In adapter_config.json: "peft.task_type" must be a string
Whisper Base Polish PEFT - s22678 prod
This model is a fine-tuned version of openai/openai/whisper-base on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5546
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: 52
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.3265 | 1.0 | 400 | 0.4049 |
| 0.2506 | 2.01 | 800 | 0.3920 |
| 0.185 | 3.01 | 1200 | 0.3868 |
| 0.1506 | 4.01 | 1600 | 0.3859 |
| 0.1217 | 5.01 | 2000 | 0.3856 |
| 0.0931 | 6.02 | 2400 | 0.3922 |
| 0.0698 | 7.02 | 2800 | 0.3999 |
| 0.0549 | 8.02 | 3200 | 0.4077 |
| 0.0477 | 9.02 | 3600 | 0.4121 |
Framework versions
- PEFT 0.11.2.dev0
- Transformers 4.36.0
- Pytorch 2.1.0
- Datasets 2.15.0
- Tokenizers 0.15.1
- Downloads last month
- 4
Model tree for MarOsz/unused-whisper-base-polish-peft-prod
Base model
openai/whisper-base