This repository contains the archives of all the zerospeech challenge leaderboard entries.
Entries are split into individual [.yaml|.json] files inside an entry folder.
For more information on the zerospeech challenges you can visit our website: zerospeech.com.
| challenge | url | status |
|---|---|---|
| zr2015 | http://zerospeech.com/tracks/2015/tasks/ | archived |
| zr2017 | http://zerospeech.com/tracks/2017/tasks/ | archived |
| zr2019 | http://zerospeech.com/tracks/2019/tasks/ | archived |
| zr2020 | http://zerospeech.com/tracks/2020/tasks/ | archived |
| zr2021 | http://zerospeech.com/tracks/2021/02_track1/ | archived |
To submit into an active challenge you are encouraged to follow the instructions on the website zerospeech.com.
Archived challenges have published benchmarks (or will be published at a future date) so evaluations of entries can happen locally, you can send us those results in the form of a pull request.
Fork this repository, add the corresponding entries into their respective folder, and create a pull request.
In your pull request it is recommended to add a link to a published paper; and explanation of your results.
extract.py : a script allowing to unpack a leaderboard.json into individual entries.
pack.py : a script allowing to pack individual entries into a single leaderboard.json file.
convert.py : a script allowing to convert yaml files to json.
If you face any issue or have questions you can open an issue on this repository, or send us an email on [email protected]