This model is a segmoe merge of 2 models from civitAI:

https://civitai.com/models/234898/vixons-fantasy-mix https://civitai.com/models/43977?modelVersionId=113623

Merged using the great project at: https://github.com/segmind/segmoe

To do something similar you can either follow the guide in readme or you can follow this blogpost: https://huggingface.co/blog/segmoe

The setting I used:

base_model: https://civitai.com/api/download/models/306781 num_experts: 4 moe_layers: all num_experts_per_tok: 2 type: sd experts:

source_model: https://civitai.com/api/download/models/306781 positive_prompt: "cinematic, portrait, photograph, instagram, fashion, movie, macro shot, 8K, RAW, fantastic, ultra high quality" negative_prompt: " (deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime), text, cropped, out of frame, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck"
source_model: https://civitai.com/api/download/models/113623 positive_prompt: "photo realistic scenes, fantastic view, impressive view, movie scene, 8K, RAW, hyperrealistic, ultra realistic" negative_prompt: "simple background, duplicate, retro style, low quality, lowest quality, 1980s, 1990s, 2000s, 2005 2006 2007 2008 2009 2010 2011 2012 2013, bad anatomy, bad proportions, extra digits, lowres, username, artist name, error, duplicate, watermark, signature, text, extra digit, fewer digits, worst quality, jpeg artifacts, blurry"

Useage

!pip install -U segmoe diffusers transformers

from segmoe import SegMoEPipeline

pipeline = SegMoEPipeline("eren23/sd15-FantasyMix-filmGrain-segmoe", device="cuda")

prompt = "fantastic land canvas, knight cat standing next to a purple medieval village wall" negative_prompt = "nsfw, bad quality, worse quality" img = pipeline( prompt=prompt, negative_prompt=negative_prompt, height=512, width=512, num_inference_steps=30, guidance_scale=7.5, ).images[0] img.save("image.png")