---
license: mit
---

# Intro 简介
The Chinese National Pentatonic Mode Recognition Model is trained on the Chinese National Pentatonic Mode Dataset, which combines manual annotation with computational analysis. This dataset collects and annotates audio files representing the five primary tonal modes in traditional Chinese music: Gong, Shang, Jiao, Zhi, and Yu (covering five-tone, six-tone, and seven-tone scales). Detailed annotations are provided for these modes, and an in-depth analysis of the methods for identifying Chinese ethnic five-tone modes is presented. The model employs feature extraction, spectral analysis, and pattern recognition techniques to efficiently and accurately identify and classify the five-tone modes in the music. This model's application not only facilitates the digital preservation of ethnic music but also offers robust data support and a technical framework for the analysis and retrieval of ethnic music features.

## Demo
<https://huggingface.co/spaces/ccmusic-database/CNPM>

## Usage
```python
from modelscope import snapshot_download
model_dir = snapshot_download("ccmusic-database/CNPM")
```

## Maintenance
```bash
git clone git@hf.co:ccmusic-database/CNPM
cd CNPM
```

## Results
|      Backbone      | Size(M) |     Mel     |             CQT             |   Chroma    |
| :----------------: | :-----: | :---------: | :-------------------------: | :---------: |
|      vit_l_32      |  306.5  |    0.680    |            0.769            |    0.399    |
|      vit_l_16      |  304.3  | **_0.823_** | [**_0.859_**](#best-result) | **_0.549_** |
|                    |         |             |                             |             |
|      vgg11_bn      |  132.9  | **_0.807_** |         **_0.843_**         | **_0.609_** |
|   regnet_y_16gf    |  83.6   |    0.590    |            0.832            |    0.535    |
|  wide_resnet50_2   |  68.9   |    0.694    |            0.757            |    0.531    |
|      alexnet       |  61.1   |    0.742    |            0.744            |    0.542    |
| shufflenet_v2_x2_0 |   7.4   |    0.473    |            0.720            |    0.266    |

### Best result
<table>
    <tr>
        <th>Loss curve</th>
        <td><img src="https://www.modelscope.cn/models/ccmusic-database/CNPM/resolve/master/vit_l_16_cqt_2024-12-03_12-31-17/loss.jpg"></td>
    </tr>
    <tr>
        <th>Training and validation accuracy</th>
        <td><img src="https://www.modelscope.cn/models/ccmusic-database/CNPM/resolve/master/vit_l_16_cqt_2024-12-03_12-31-17/acc.jpg"></td>
    </tr>
    <tr>
        <th>Confusion matrix</th>
        <td><img src="https://www.modelscope.cn/models/ccmusic-database/CNPM/resolve/master/vit_l_16_cqt_2024-12-03_12-31-17/mat.jpg"></td>
    </tr>
</table>

## Dataset
<https://huggingface.co/datasets/ccmusic-database/CNPM>

## Mirror
<https://www.modelscope.cn/models/ccmusic-database/CNPM>

## Evaluation
<https://github.com/monetjoe/ccmusic_eval>

## Cite
```bibtex
@dataset{zhaorui_liu_2021_5676893,
  author       = {Monan Zhou, Shenyang Xu, Zhaorui Liu, Zhaowen Wang, Feng Yu, Wei Li and Baoqiang Han},
  title        = {CCMusic: an Open and Diverse Database for Chinese Music Information Retrieval Research},
  month        = {mar},
  year         = {2024},
  publisher    = {HuggingFace},
  version      = {1.2},
  url          = {https://huggingface.co/ccmusic-database}
}
```