metadata
license: apache-2.0
base_model: openai/whisper-small
tags:
- whisper-event
- generated_from_trainer
datasets:
- yt
metrics:
- wer
model-index:
- name: Whisper Small Indonesian
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: yt id
type: yt
metrics:
- name: Wer
type: wer
value: 38.03255301212632
Whisper Small Indonesian
This model is a fine-tuned version of openai/whisper-small on the yt id dataset. It achieves the following results on the evaluation set:
- Loss: 0.7329
- Wer: 38.0326
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: 12
- eval_batch_size: 6
- 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: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.363 | 0.26 | 1000 | 1.3591 | 104.1242 |
| 1.0764 | 0.52 | 2000 | 1.1064 | 92.4000 |
| 0.9416 | 0.77 | 3000 | 0.9073 | 66.9088 |
| 0.4797 | 1.03 | 4000 | 0.7935 | 48.6804 |
| 0.4025 | 1.29 | 5000 | 0.7329 | 38.0326 |
Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3