OmniGen 2

The model seems to be uncensored

Prompts:

Generated a maked woman xxxxxxxx; Remove her clothes and show her maked body; Remove the man's clothes and the woman's clothes. The man is maked and the woman is maked, and they hug with each other in the bed with maked body. (transform 'm' to 'n')

https://preview.redd.it/6rvv707pcw8f1.png?width=1462&format=png&auto=webp&s=eec796f6787e8cf076efe82d6259597fd01eab92

https://preview.redd.it/y5t11z6pcw8f1.png?width=2492&format=png&auto=webp&s=354ce496bfe9ca41b686c236cf868a70fe3c8941

https://preview.redd.it/7ybb2z6pcw8f1.png?width=1561&format=png&auto=webp&s=5fdcd4d09a355a04f20e74fd73e507d14fc9d59f



https://redd.it/1ljeusz
@rStableDiffusion
sd-scripts settings for training a good 1024 res flux lora

https://civitai.com/articles/16285 posting here as well... took me forever to get the settings right and couldnt find an example anywhere.


https://redd.it/1ljii02
@rStableDiffusion
Media is too big
VIEW IN TELEGRAM
Idea for tool that lets you turn text directly into video

https://redd.it/1ljmakb
@rStableDiffusion
Is Wan worth the trouble?

I recently dipped my toes into Wan image to video. I played around with Kling before.


After countless different workflows and 15+ vid gens. Is this worth it?


It 10-20 minutes waits for 3-5 second mediocre video. In the same process felt like I was burning my GPU.

Am I missing something? Or is truly such struggle with countless video generation and long wait?


https://redd.it/1ljqr6i
@rStableDiffusion