metadata
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- lyhourt/clean_6
metrics:
- wer
model-index:
- name: whisper-small-clean_6-v6
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: lyhourt/clean_6
type: lyhourt/clean_6
metrics:
- name: Wer
type: wer
value: 26.64601084430674
whisper-small-clean_6-v6
This model is a fine-tuned version of openai/whisper-small on the lyhourt/clean_6 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3768
- Wer: 26.6460
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: 64
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.1354 | 0.2 | 100 | 0.4427 | 34.7792 |
| 0.1115 | 0.4 | 200 | 0.3909 | 27.7304 |
| 0.1105 | 1.002 | 300 | 0.3762 | 27.8854 |
| 0.0392 | 1.202 | 400 | 0.3856 | 26.6460 |
| 0.0444 | 1.4020 | 500 | 0.3768 | 26.6460 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1