Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
task_categories:
|
| 4 |
+
- automatic-speech-recognition
|
| 5 |
+
language:
|
| 6 |
+
- zh
|
| 7 |
+
size_categories:
|
| 8 |
+
- n<1K
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Test Audio Dataset
|
| 12 |
+
|
| 13 |
+
这是一个用于测试的音频数据集，包含 100 条伪造的 WAV 格式音频文件。
|
| 14 |
+
|
| 15 |
+
## Dataset Description
|
| 16 |
+
|
| 17 |
+
- **语言**: 中文 (zh)
|
| 18 |
+
- **音频格式**: WAV (16kHz, 16-bit, 单声道)
|
| 19 |
+
- **样本数量**: 100
|
| 20 |
+
- **用途**: 测试语音识别/合成流程
|
| 21 |
+
|
| 22 |
+
## Dataset Structure
|
| 23 |
+
|
| 24 |
+
```
|
| 25 |
+
audio-dataset/
|
| 26 |
+
├── data/
|
| 27 |
+
│ ├── audio_0000.wav
|
| 28 |
+
│ ├── audio_0001.wav
|
| 29 |
+
│ └── ...
|
| 30 |
+
├── metadata.csv
|
| 31 |
+
└── README.md
|
| 32 |
+
```
|
| 33 |
+
|
| 34 |
+
## Data Fields
|
| 35 |
+
|
| 36 |
+
- `file_name`: 音频文件路径
|
| 37 |
+
- `transcription`: 转录文本
|
| 38 |
+
- `speaker_id`: 说话人ID
|
| 39 |
+
- `duration`: 音频时长（秒）
|
| 40 |
+
- `sample_rate`: 采样率
|
| 41 |
+
|
| 42 |
+
## Usage
|
| 43 |
+
|
| 44 |
+
```python
|
| 45 |
+
from datasets import load_dataset
|
| 46 |
+
|
| 47 |
+
dataset = load_dataset("your-username/test-audio-dataset")
|
| 48 |
+
```
|
| 49 |
+
|
| 50 |
+
## License
|
| 51 |
+
|
| 52 |
+
MIT License
|