Hello,
Is it allowed to upload model weights on the HuggingFace platform after training on the TalkBank corpus dataset?
What kind of license should be put? How can I refer to the TalkBank corpus dataset on the model card?
Can I hide model weights from others and open access after the competition deadline? I will put the HF API token in the code if required.
Thank you in advance.
Hi @musayev - thanks for the thoughtful questions!
You may publicly share model weights developed for this competition during the competition period or after the deadline under an MIT license. Reminder that, per the rules, “privately sharing code or data outside of teams is not permitted.”
Note that you are responsible for ensuring adherence to the challenge rules, including never sharing or republishing the challenge data.
We encourage you to showcase your model using the “Share your work” feature on the challenge website (the link will appear next to your submission on the leaderboard after the competition ends).
Required attribution information is listed on the About page. If you share your model during the competition, please use the attribution information currently provided and update your model card after the competition deadline when the full citation list is posted.