metadata
language:
- en
license: apache-2.0
base_model: openai/whisper-small.en
tags:
- hf-asr-leaderboard
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper Base EN
results: []
Whisper Base EN
This model is a fine-tuned version of openai/whisper-small.en on the ADLINK dataset. It achieves the following results on the evaluation set:
- Loss: 0.0002
- Wer: 1.5152
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.1465 | 25.0 | 100 | 1.1124 | 9.6970 |
| 0.4228 | 50.0 | 200 | 0.4547 | 2.4242 |
| 0.0555 | 75.0 | 300 | 0.0459 | 1.8182 |
| 0.0022 | 100.0 | 400 | 0.0022 | 1.8182 |
| 0.0007 | 125.0 | 500 | 0.0008 | 1.5152 |
| 0.0004 | 150.0 | 600 | 0.0005 | 1.5152 |
| 0.0003 | 175.0 | 700 | 0.0003 | 1.5152 |
| 0.0002 | 200.0 | 800 | 0.0003 | 1.5152 |
| 0.0002 | 225.0 | 900 | 0.0003 | 1.5152 |
| 0.0002 | 250.0 | 1000 | 0.0002 | 1.5152 |
Framework versions
- Transformers 4.41.0
- Pytorch 2.2.0a0+81ea7a4
- Datasets 2.19.1
- Tokenizers 0.19.1