search

lixirui142 / VidToMe

Official Pytorch Implementation for "VidToMe: Video Token Merging for Zero-Shot Video Editing" (CVPR 2024)
https://vidtome-diffusion.github.io
MIT License
173 stars 10 forks source link

HF integration #5

Open jadechoghari opened 2 weeks ago

jadechoghari commented 2 weeks ago

Hello @lixirui142 and team, Congratulations on this amazing work—incredible! 🎉 🚀

You should consider making an native implementation of VidToMe to HF hub 🤗. It can easily be made compatible natively with the Hugging Face Transformers/Diffusers library.

Will do the work - other folks from HF will be available to help. From your behalf if any questions arise during the process we can open an issue :)

Soon, it will be easily usable with just one line of code. And nice gradio demo on spaces!!

Best regards, Jade

lixirui142 commented 1 week ago

Hi Jade,

Thank you so much for the kind words and your interest in VidToMe! 🎉🚀 I'm excited about the idea of integrating it with the Hugging Face ecosystem. I will definitely give it a try when I have some time.

In the meantime, I wonder whether there are guidelines for implementation. That would be really helpful as I explore how best to approach this.

jadechoghari commented 1 week ago

@lixirui142 Actually: "Will do the work - other folks from HF will be available to help.", so you don't have to worry about that. We'll send you an email at lixirui142@sjtu.edu.cn :) for further instructions!