---
language:
- zh
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
datasets:
- formospeech/tat_asr_aligned
metrics:
- wer
model-index:
- name: Whisper Tiny Taiwanese Condenser
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: TAT ASR Aligned
      type: formospeech/tat_asr_aligned
      args: 'config: taiwanese, split: test'
    metrics:
    - name: Wer
      type: wer
      value: 50.65108143376739
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Whisper Tiny Taiwanese Condenser

This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the TAT ASR Aligned dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0814
- Wer: 50.6511

## 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: 0.001
- train_batch_size: 64
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1362
- training_steps: 13620
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch   | Step  | Validation Loss | Wer     |
|:-------------:|:-------:|:-----:|:---------------:|:-------:|
| 0.4229        | 0.9985  | 681   | 0.6344          | 65.8772 |
| 0.4529        | 1.9971  | 1362  | 0.7311          | 71.3135 |
| 0.3271        | 2.9956  | 2043  | 0.6324          | 64.7777 |
| 0.2257        | 3.9941  | 2724  | 0.5510          | 58.4141 |
| 0.1554        | 4.9927  | 3405  | 0.5350          | 57.6394 |
| 0.122         | 5.9912  | 4086  | 0.5911          | 56.8148 |
| 0.0935        | 6.9897  | 4767  | 0.6122          | 56.1179 |
| 0.0718        | 7.9883  | 5448  | 0.6492          | 54.7158 |
| 0.0588        | 8.9868  | 6129  | 0.6623          | 55.5599 |
| 0.0466        | 9.9853  | 6810  | 0.6883          | 56.9481 |
| 0.0349        | 10.9839 | 7491  | 0.7069          | 54.6770 |
| 0.0298        | 11.9824 | 8172  | 0.7441          | 54.2272 |
| 0.0215        | 12.9809 | 8853  | 0.7937          | 54.9491 |
| 0.0141        | 13.9795 | 9534  | 0.8062          | 52.9278 |
| 0.01          | 14.9780 | 10215 | 0.8717          | 53.2804 |
| 0.0067        | 15.9765 | 10896 | 0.9279          | 52.6029 |
| 0.0031        | 16.9751 | 11577 | 0.9783          | 52.0393 |
| 0.0012        | 17.9736 | 12258 | 1.0311          | 50.6622 |
| 0.0004        | 18.9721 | 12939 | 1.0574          | 50.7566 |
| 0.0001        | 19.9707 | 13620 | 1.0814          | 50.6511 |


### Framework versions

- Transformers 4.42.3
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1