Simp Chat Emma Watson DeepFakes AI prompts

[Slava 11]

Hello guys!

I think it's time to make a separate thread on the topic of successful and working prompts for generating deepfakes with Emma. Together with several forum members, we discussed various techniques, this happened in private messages. Perhaps it would be useful to learn together and share experience in a separate thread

The neural networks that currently work most effectively — Reve, Vertex AI, Grok, and a few others — work with video.


In this thread you can:

send photo ideas/references for generation;
share successful prompts that the neural system misses;
talk about new tools, life hacks and auxiliary tools, such as for improving photo quality.

 

[Slava 11]

roman1205 pull4boi99 KazzFakes Simp730 TheGalGadotMatrix, join the conversation

 

[Slava 11]

Yesterday at 4 am, while doing this "charitable and creative work" , I discovered that with the "right angle" of selection in Adobe Photoshop, their Generative Fill tool can remove lingerie from generated images from Reve. If you add to this the sharpness effect and 5% grain, as in film photography, it creates an imitation of real shooting.

The biggest challenge at the moment is working with hands and feet, which all neurons have trouble with. Sometimes I correct the eyes if they come out square, not round (the iris).

 

[Slava 11]

References above I gonna try next.

How it works? I send it to Grok or ChatGPT and ask it to describe an image as a prompt but change the unknown model with Emma. Then – I generate images on Reve with it and changing it during process.


I

 

[TheGalGadotMatrix]

References above I gonna try next.

How it works? I send it to Grok or ChatGPT and ask it to describe an image as a prompt but change the unknown model with Emma. Then – I generate images on Reve with it and changing it during process.


I

That's a genius way to get prompts and afterward to tweak it however you want to but the baseline idea is written out by Grok.