---
title: FusionAudio-6
canonical_url: "https://www.modelscope.cn/datasets/FreedomIntelligence/FusionAudio-6"
md_url: "https://www.modelscope.cn/datasets/FreedomIntelligence/FusionAudio-6.md"
repository: FreedomIntelligence/FusionAudio-6
last_updated: 2025-06-22
license: "Apache License 2.0"
storage_size: "132 GB"
downloads: 2077
stars: 0
---

# FusionAudio-6

> FusionAudio-6 - FreedomIntelligence 在 ModelScope 开源的数据集。Audio Caption Dataset

FreedomIntelligence/FusionAudio-6 是 ModelScope 魔搭社区上的数据集，存储大小 132 GB，采用 Apache License 2.0 许可。

- **Repository**: FreedomIntelligence/FusionAudio-6
- **License**: Apache License 2.0
- **Storage size**: 132 GB
- **Downloads**: 2077
- **Stars**: 0
- **Last updated**: 2025-06-22

Source: https://www.modelscope.cn/datasets/FreedomIntelligence/FusionAudio-6

---

# Audio Caption Dataset

## Dataset Description

This dataset contains audio captions from the dataset, including:
- Audio captions
- Speech captions  
- Music captions (where available)
- Audio files and metadata

## Dataset Structure

The dataset contains the following columns:
- `id`: Unique identifier for each audio sample
- `audio_caption`: Description of the audio content
- `speech_caption`: Description of speech content
- `music_caption`: Description of music content (may be empty)
- `path`: Relative path to the audio file
- `url`: Original URL (if available)
- `timeinterval`: Time interval information
- `audio`: Audio data

## Usage

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("tzzte/FusionAudio-6")

# Access a sample
sample = dataset["train"][0]
print(sample["audio_caption"])
```