Stable Diffusion, NovelAI, Machine Learning Art - AI art generation discussion and image dump

Flux2 img2img is pretty good. I gave it a picture of myself and told it to make the person more attractive, and it returned the exact same picture except it removed some snow from the background. The model knows its stuff.
 
Crash Bandicoot visiting the wailing wall
crash the wall.png

Made with Gemini
 
Not quite--he should be punching the wall with an angry look on his face.

"Grandpa, stop--you're embarrassing us!"
"I-I told you what would happen if you made me come here, Morty--you know that I have VERY STRONG, VERY REASONABLE FEELINGS about the state of the whole GODDAMN FUCKING SHITSHOW that these people have involved themselves in for thousands of years AND CONTINUE TO DO SO..."
 
Last edited:
Not quite--he should be punching the wall with an angry look on his face.

"Grandpa, stop--you're embarrassing us!"
"I-I told you what would happen if you made me come here, Morty--you know that I have VERY STRONG, VERY REASONABLE FEELINGS about the state of the whole GODDAMN FUCKING SHITSHOW that these people have involved themselves in for thousands of years AND CONTINUE TO DO SO..."
you know what
You're correct
oy rage 1.pngoy rage 2.png
 
> update and restart comfyui
> "a new update is available"
> update and restart comfyui again
> "a new update is available"


Nigger how
Dependencies upon dependencies upon dependencies and custom nodes. I hadn't updated ComfyUI in six months or so and needed to restart at least three times before everything was up to date.
 
> update and restart comfyui again
> every workflow you use is now broken
This is why I haven't moved on from Forge Neo even if there's some interesting workflows I've been wanting to try. I use Krita's AI plugin that uses ComfyUI in the backend and fuck me if something doesn't break every update. Mashallah save me from the spaget.
 
This is why I haven't moved on from Forge Neo even if there's some interesting workflows I've been wanting to try. I use Krita's AI plugin that uses ComfyUI in the backend and fuck me if something doesn't break every update. Mashallah save me from the spaget.
It's a nightmare. Every time I get my main workflows working the way I want I have to make a backup of my comfy install before I install any updates because I've been fucked over multiple times.
 
downloaded LTX 2.3 and started playing around and it's pretty good.


 
The video extender option of LTX 2.3 is scarily good.

you can even do copyrighted characters with the base model with text to video.
 
Hey, could y'all LTX users do me a favor and generate stuff like commercials for pajeet off or corporate meetings/infomercially style prompts? I'm seriously considering doing the following quote and think this might be da way to da queen.

This thread made me MATI, I think I'm gonna make an AI ARG.
Please n' thank you.
 
Slowly coming to a workflow that works, at least now I got NTSC-RS working via a plug in Davinci Resolve. Only had about 15 minutes, I think there is potential just gotta figure out what I need to do with the raw AI clip to get audio and Davinci talking to one another.

 
The Bing gen AI died, stuck in a loop of needing to login. Any other free online txt2img ones people would recommend? It's been pretty useful for TTRPG stuff
 
Back
Top Bottom