search

baaivision / Emu

Emu Series: Generative Multimodal Models from BAAI
https://baaivision.github.io/emu2/
Apache License 2.0
1.66k stars 86 forks source link

Image blending prompt #2

Closed vishaal27 closed 1 year ago

vishaal27 commented 1 year ago

Hey, thanks so much for releasing your great paper and making the code and weights public. I am quite interested in playing around with the model for the task of image blending. Could you please provide the exact prompt used for the image blending results in Figure 1 of the paper?

yqy2001 commented 1 year ago

No prompt is needed for image blending :). To blend two images, simply concatenate them into a sequence, and append an [IMG] token to denote the start of the image to be generated:

$$ \texttt{[IMG]image1[/IMG][IMG]image2[/IMG][IMG]} $$

, then feed this sequence into pretrained Emu, which will auto-regressively generates 32 image embeddings that could be further decoded into realistic blended image.

Please note that we have not released the code for auto-regressive image generation and the weights of the image decoder. However, you can play with our demo now, which supports this function.

Thank you for your interest in our work.

PeihaoChen commented 1 year ago

Thanks for your amazing work! When will you release the code for auto-regressive image generation and the weights of the image decoder? Also, do you have a plan to release the training code? Thx!

yqy2001 commented 1 year ago

We plan to release all of them, but it might take some time to do so.

We appreciate your interest and please wait for our release :). Thank you.