metadata
language:
- ta
license: apache-2.0
base_model: openai/whisper-medium
tags:
- generated_from_trainer
datasets:
- tamilcustomvoice
metrics:
- wer
model-index:
- name: Whisper tiny custom
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: custom dataset
type: tamilcustomvoice
metrics:
- name: Wer
type: wer
value: 7.28476821192053
Whisper tiny custom
This model is a fine-tuned version of openai/whisper-medium on the custom dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.0315
- Wer Ortho: 9.2105
- Wer: 7.2848
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: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 1.6536 | 2.5 | 50 | 0.4681 | 57.8947 | 50.9934 |
| 0.0732 | 5.0 | 100 | 0.0820 | 19.7368 | 15.2318 |
| 0.0076 | 7.5 | 150 | 0.0396 | 9.2105 | 7.9470 |
| 0.0013 | 10.0 | 200 | 0.0336 | 9.2105 | 8.6093 |
| 0.0007 | 12.5 | 250 | 0.0356 | 7.8947 | 5.9603 |
| 0.0005 | 15.0 | 300 | 0.0339 | 7.8947 | 5.9603 |
| 0.0004 | 17.5 | 350 | 0.0326 | 7.8947 | 5.9603 |
| 0.0003 | 20.0 | 400 | 0.0323 | 7.8947 | 5.9603 |
| 0.0003 | 22.5 | 450 | 0.0320 | 9.2105 | 7.2848 |
| 0.0002 | 25.0 | 500 | 0.0315 | 9.2105 | 7.2848 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1