Playground V2.5 with DMD2 Refiner (Face Detailer, Ollama, Siax upscale).

It's pretty simple, an image is made (with Dynamic Prompting and Ollama if you want to enhance the prompt) with Playground V2.5, which is always a bit messy, it goes to a Face Detailer in case you need it, then is uspcaled to 2 Megapixels, then refined by DMD2 at 20-25%, that cleans up the mess.

There's a group muter in case you want to bypass something, and a node that strips the <think></think> part of the prompt in case you use DeepSeeker.

Version 2.0 is refined, with bigger upscale, film grain and better organisation.