search

Junyi42 / sd-dino

Official Implementation of paper "A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence"
https://sd-complements-dino.github.io
255 stars 12 forks source link

License? #13

Open iago-suarez opened 6 months ago

iago-suarez commented 6 months ago

Hi,

Thanks for this awesome work! 🤩

DINO and StableDiffusion works have MIT licenses. Is your work also MIT?

Best, Iago.

Junyi42 commented 6 months ago

Hi Iago,

Thanks for your interest in our work!

We intend to release it under MIT, but unfortunately, we use the implicit text embedding from ODISE when extracting the SD features, which is under NVIDIA's license. We are planning to remove this part in the future and release under MIT license.

Currently, you could use the DIFT implementation for SD feature extractor to replace the ODISE's version in https://github.com/Junyi42/sd-dino/blob/master/extractor_sd.py, and this way will be under MIT.

Best, Junyi