OpenGVLab/DragGAN · Apply for community grant: Academic project

Greeting! A few days ago, we discovered a very good work: DragGAN, which can use keypoints for control and achieve high degrees of freedom and fidelity in image editing, including shape, posture and expression. We found this work very interesting and reproduced it.

We built a visual demo using Gradio for everyone to experience, and our code is open source at https://github.com/Zeqiang-Lai/DragGAN. It has now received 2k stars! We would love to share this work with huggingface!

As a non-profit institution dedicated to the development and open-source of general vision models, OpenGVLab would like to extend our demo to host our demo on Huggingface with GPU support. Due to budget constraints, we are unable to cover GPU expenses ourselves. Therefore, we humbly request that you consider our demo requirements, which call for a GPU with 12GB or larger vRAM. Thank you for considering our proposal.