metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-tiny-khmer-aug-v6
results: []
whisper-tiny-khmer-aug-v6
This model is a fine-tuned version of openai/whisper-tiny on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1860
- Wer: 100.0
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: 0.0001
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.7 | 0.9994 | 830 | 0.3232 | 100.0 |
| 0.2937 | 2.0 | 1661 | 0.2460 | 100.0 |
| 0.2284 | 2.9994 | 2491 | 0.2058 | 100.0 |
| 0.1897 | 4.0 | 3322 | 0.1920 | 100.0 |
| 0.1645 | 4.9994 | 4152 | 0.1828 | 100.0 |
| 0.1455 | 6.0 | 4983 | 0.1818 | 100.0 |
| 0.1315 | 6.9994 | 5813 | 0.1781 | 100.0 |
| 0.1189 | 8.0 | 6644 | 0.1829 | 100.0 |
| 0.1089 | 8.9994 | 7474 | 0.1843 | 100.0 |
| 0.0988 | 9.9940 | 8300 | 0.1860 | 100.0 |
Framework versions
- Transformers 4.44.0
- Pytorch 2.3.1
- Datasets 2.21.0
- Tokenizers 0.19.1