search

johko / computer-vision-course

This repo is the homebase of a community driven course on Computer Vision with Neural Networks. Feel free to join us on the Hugging Face discord: hf.co/join/discord
MIT License
452 stars 139 forks source link

Unit 3 - DINAT Transformer #177

Closed alanahmet closed 8 months ago

alanahmet commented 8 months ago

Hi everyone,

This is a pull request for Unit 3 - DINAT Transformer. I apologize for the delay PR. I was facing some health issues. I hope it will be fine

alanahmet commented 8 months ago

Great section, thanks a lot 🫡 To be perfectly candid, I think OneFormer deserves it's own page. I've already written about it: https://x.com/mervenoyann/status/1739707076501221608?s=20 if you feel like it, you can merge those and create a new section for OneFormer and just give a link to that section from this section saying that DiNAT is used in OneFormer. WDYT?

Thanks Merve DINAT seems to be the most suitable architecture for the former model as indicated in the paper. That's why I included it here. However, it would make more sense to create a separate section. Which one would be better creating a Jupyter Notebook or an mdx file for oneformer?

merveenoyan commented 8 months ago

@alanahmet mdx file works! 🤝 thank you 💜