- This script for downloading AVSpeech Dataset.
- It downloads only the relevant part of the video and audio.
- The script creates a new folder named 'data' under the working directory.
- Each downloaded audio and video is placed in a folder named after the video id
- Name of folder
yt_id - Name of video and audio files
yt_id\_start_time\_end_time.mp4/mp3 - The script creates a file called
bad_files_csv.txtwhich lists the youtube id's of the deleted/private videos which are no longer available for download.
- Name of folder
Usage:
pip install -r requirements.txt
python downloader.py --csv ./example.csv
Features:
| Features | Required | Values |
|---|---|---|
| --csv | ✅ | csv path |
| --format | ❌ | best,worst,22/18,137/140 |
| --audio | ❌ | True/False |
'best' = The best quality of video 'worst' = The worst quality of video '22/18' = 720p '137/140' = 1080p
NOTE : You can change the number of threads according to your resource constraints. Happy downloading!