Tabahi/yt-cc-speakers-6spk

Tasks

• automatic-speech-recognition
• text-to-speech

Citation text from the card

The copyright in each recording belongs to the creator of the source video,
not to the dataset maintainers. `ATTRIBUTION.csv` gives the credit line for
every video, one row per recording:

| Column | Content |
|---|---|
| `split` | `train` or `test` |
| `spkr` | Speaker ID used in this dataset |
| `videoId` | YouTube video ID |
| `title` | Title of the source video |
| `creator` | YouTube channel that published the video (the author to credit) |
| `source_url` | Link to the source video |
| `channel_url` | Link to the creator's channel |
| `published_at` | Publication date of the source video |
| `license`, `license_url` | CC BY 3.0 and the link to the license |
| `audio`, `captions` | The files in this dataset derived from that video |

If you redistribute any of this data, or publish work that reproduces it, you
must:

1. Credit the creator and give the title of the video.
2. Link to the source video.
3. Name the license (CC BY 3.0) and link to it.
4. State that changes were made (see below).
5. Not suggest that the creator endorses you or your use.
6. Not add legal terms or technical measures that stop others from doing what
   the license allows.

A credit in this form satisfies points 1 to 4:

> "VLOG: chill day at home, my new coworker & pumpkin cold brew is back" by Kenzie Elizabeth (https://www.youtube.com/watch?v=f9s7NR4MAjs), licensed under CC BY 3.0
> (https://creativecommons.org/licenses/by/3.0/). Audio extracted, converted
> to 16 kHz mono, and denoised; transcri

All datasets marked Permissive or share-alike | JSON