ComfyUI Tutorial: Create Two Talking AI Characters On 6GB VRAM
https://youtu.be/DAtAp4HlErE

https://redd.it/1tus405
@rStableDiffusion

Fizgig Klein 9b Lora Studio v1.2.4 - update targeting 16gb Card users
https://redd.it/1tuq8lw
@rStableDiffusion

PixelDiT: Pixel Diffusion Transformers
for Image Generation
Pixel Diffusion Transformers
for Image Generation, 1.3B, no VAE
https://pixeldit.github.io/

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

Anima testing for complex scene
https://redd.it/1tuy3ye
@rStableDiffusion

MISO-TTS . 8 Billion text2speech model released.

https://redd.it/1tux5qx
@rStableDiffusion

Do you listen to your GPU?

After watching a movie/show with the kids, my PC gets switched back to my own monitor and the regular sound comes out of my speakers again, but the main speakers are still hooked up, and they are not silent.

It's like a cross between 1980's cassette-loading software and Alva Noto, and I admit, I often leave the main hifi switched to the PC channel just to listen to this in the background. It can get truly musical, sometimes with definite beats, basslines, and more. There are other ways to get this noise.

Anyway, I like to listen to my GPU working and wondered if I was alone in this.

https://redd.it/1tv3vzt
@rStableDiffusion

I compared 62 samplers and 16 schedulers for Z-Image Turbo and rated the image quality so you don't have to 😉

https://preview.redd.it/vvu48gf14y4h1.png?width=616&format=png&auto=webp&s=5ea23d0687e6d27682afc399f7c3f577ed15aa40

Here's a sampler/scheduler comparison table for image generation with Z-Image Turbo. Obviously it reads like Red < Orange < Yellow < Green. You're welcome!

PS. If you don't like it, don't appreciate it or think I'm wasting my time... Then... Don't waste your time, just move along 😉

https://redd.it/1tv6et1
@rStableDiffusion

UPDATE NexusBTA v0.2.22 is out Ui with pre made Comfy Workflows

https://redd.it/1tv7fii
@rStableDiffusion

Some Cosmic Fantasy Generations with Anima (Prompts Included)

https://redd.it/1tv6o42
@rStableDiffusion

Geometrically consistent 360-degree scenes from single panoramas

https://redd.it/1tv6l3c
@rStableDiffusion

Bytedance Bernini workflow

https://redd.it/1tv4xwf
@rStableDiffusion

Why do people like flux2 klein edit so much?

Basically the title. I've played around with the edit functionality a fair bit, and it just doesn't seem that good compared to qwen image edit. It changes the lighting, distorts faces, and gives weird anatomy or composition randomly. It's nice and fast, but accuracy and quality don't seem that good.

What am I missing?

https://redd.it/1tvb2pa
@rStableDiffusion

More person to DnD

https://redd.it/1tvb0ff
@rStableDiffusion