metadata
library_name: transformers
language:
- en
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper Small
results: []
Whisper Small
This model is a fine-tuned version of openai/whisper-small on the Personal - Mimic Recording dataset. It achieves the following results on the evaluation set:
- Loss: 0.1889
- Wer: 0.0811
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: 2e-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: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.6843 | 1.0 | 74 | 0.3695 | 0.1695 |
| 0.2675 | 2.0 | 148 | 0.2706 | 0.1350 |
| 0.1288 | 3.0 | 222 | 0.2188 | 0.1038 |
| 0.0443 | 4.0 | 296 | 0.1949 | 0.0983 |
| 0.0165 | 5.0 | 370 | 0.1943 | 0.0898 |
| 0.0129 | 6.0 | 444 | 0.2140 | 0.0932 |
| 0.0111 | 7.0 | 518 | 0.1999 | 0.0898 |
| 0.0069 | 8.0 | 592 | 0.1927 | 0.0922 |
| 0.0038 | 9.0 | 666 | 0.1904 | 0.0758 |
| 0.0021 | 10.0 | 740 | 0.1889 | 0.0811 |
Framework versions
- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0