metadata
license: apache-2.0
base_model: openai/whisper-medium
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-medium-studio-records
results: []
whisper-medium-studio-records
This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0715
- Wer: 37.1734
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: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 51000
- training_steps: 6000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3749 | 0.41 | 1000 | 0.4543 | 87.2605 |
| 0.1073 | 0.82 | 2000 | 0.1552 | 62.8266 |
| 0.0705 | 1.23 | 3000 | 0.1148 | 52.0527 |
| 0.051 | 1.64 | 4000 | 0.0935 | 45.3098 |
| 0.0381 | 2.06 | 5000 | 0.0801 | 41.0550 |
| 0.0336 | 2.47 | 6000 | 0.0715 | 37.1734 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.1+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1