metadata
language:
- he
license: apache-2.0
base_model: openai/whisper-small
tags:
- hf-asr-leaderboard
- generated_from_trainer
metrics:
- wer
model-index:
- name: he
results: []
he
This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.7915
- Wer: 208.7179
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 6
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 2.3845 | 0.05 | 1 | 2.3331 | 122.4359 |
| 2.3669 | 0.09 | 2 | 2.3331 | 122.4359 |
| 2.327 | 0.14 | 3 | 2.3331 | 122.4359 |
| 2.3936 | 0.18 | 4 | 2.0481 | 115.1282 |
| 2.0786 | 0.23 | 5 | 1.8996 | 229.3590 |
| 1.9254 | 0.27 | 6 | 1.7915 | 208.7179 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1