ODSS / README.md
Yixuan's picture
Update README with data preprocessing instructions
9ed83da verified
metadata
license: apache-2.0

ODSS 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_odss.sh
./download/get_odss.sh

Label Distribution

The dataset contains 30,025 samples:

  • spoof: 18,993 samples
  • bonafide: 11,032 samples

Data Attributes

                                   ID                                                          path     label dataset_name
0  atlanticnarratives2_11_thomas_0034  natural/hifi-tts/6670/atlanticnarratives2_11_thomas_0034.wav  bonafide         ODSS
1          buffcollie_08_terhune_0036          natural/hifi-tts/6670/buffcollie_08_terhune_0036.wav  bonafide         ODSS
2           portent_16_macdonald_0087           natural/hifi-tts/6670/portent_16_macdonald_0087.wav  bonafide         ODSS

How to Import

import pandas as pd

df = pd.read_parquet("odss.parquet")
print(df.head())