---
license: apache-2.0
---

# SpoofCeleb Dataset

Please use the download scripts from https://github.com/XIAOYixuan/AUDDT/tree/yixuan-dev
to download and process the dataset.

```
chmod +x download/get_spoofceleb.sh
./download/get_spoofceleb.sh
```

## Label Distribution

**train** (2,540,421 samples):
- spoof: 2,309,473 samples
- bonafide: 230,948 samples

**development** (55,741 samples):
- spoof: 47,778 samples
- bonafide: 7,963 samples

**evaluation** (91,130 samples):
- spoof: 82,017 samples
- bonafide: 9,113 samples

## Data Attributes

```
  ID                                                          path     label dataset_name
0  1  spoofceleb/flac/train/a00/id11178/LAHad7VLKXw-00007-001.flac  bonafide   SpoofCeleb
1  2  spoofceleb/flac/train/a00/id10580/VR8MUnB_MTI-00007-003.flac  bonafide   SpoofCeleb
2  3  spoofceleb/flac/train/a00/id10105/vJSc3FB4US0-00067-002.flac  bonafide   SpoofCeleb
```

## How to Import

```python
import pandas as pd

df = pd.read_parquet("train.parquet")
print(df.head())
```