Apply for community grant: Academic project

#1
by User1342 - opened

WatchTower is a Twitter tool for identifying and removing violent, hate speech, and extremist content from your feed. It's derived from my PhD research into Ideologically Motivated Cyber Crime and is designed to help people curate their feeds. The web app is currently backed by two machine learning models, for identifying hate speech and extremism, with more planned for the future.

Thanks for your consideration,
James

Hi James, your demo looks amazing. I think the login with Twitter didn't work from within the Spaces though - not sure if it's an issue on our end or you are still working on it;

Also, I see you are hosting the ML models in the repo here itself: maybe is worth considering uploading the weights of the models to the Hub to help other people use those models as well 🤗, let me know if I can be more helpful

Hey, thanks!
Yeah, it's a known issue, due to how the URL params are handled (see here: https://discuss.huggingface.co/t/hugging-face-and-gradio-url-paramiters/21110/3), due to that it primarily only works from the embed view currently, here: https://user1342-watchtower.hf.space/.

Good point with the models! I'll look into that this week.

User1342 changed discussion status to closed
User1342 changed discussion status to open

Sign up or log in to comment