---
language:
- zh
license: apache-2.0
base_model: ZhihCheng/whisper-tiny-zh
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper tiny ZH
  results: []
datasets:
- mozilla-foundation/common_voice_16_1
---

<!-- 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 ZH

This model is a fine-tuned version of [ZhihCheng/whisper-tiny-zh](https://huggingface.co/ZhihCheng/whisper-tiny-zh) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4593
- Wer: 67.1021

## 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: 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: 4000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer     |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.1481        | 1.32  | 1000 | 0.4392          | 67.0820 |
| 0.0898        | 2.65  | 2000 | 0.4387          | 66.4188 |
| 0.0603        | 3.97  | 3000 | 0.4486          | 66.7002 |
| 0.0273        | 5.3   | 4000 | 0.4593          | 67.1021 |


### Framework versions

- Transformers 4.38.0.dev0
- Pytorch 2.2.0a0+6a974be
- Datasets 2.16.1
- Tokenizers 0.15.1