Images are generated without LoRAs, using detailers. Example prompts can be found in the preview images.

Positive Prompt (optional, usually not required):
masterpiece, amazing quality, best quality, ultra-detailed

Negative Prompt (recommended):
lowres, bad quality, worst quality

Negative Prompt (extended version, optional):
lowres, bad quality, worst quality, bad anatomy, jpeg artifacts, ugly, poorly drawn, blurry, censored, artist name, signature, watermark

Example of prompt tags that affect generation visual style:
realistic, photorealistic, toon style, anime style, kawaii style, illustration style, retro style, fantasy style, dark fantasy style, 1980s style, clipart style, comic style, vaporwave style, synthwave style, cel shading style, dreamcore style, manga style, dark academia style, light academia style, steampunk style, cyberpunk style, hentai style, surreal style, abstract style

The effect is more pronounced for versions V1, V2 and V3.

Recommended Settings

Steps: 25-40
CFG Scale: 4-7
Sampler: Euler a / DPM++ 2M
Scheduler: Karras
Recommended Resolution:
768×1344, 832×1216, 768×1280, 704×1408, 1024×1536

V5 merge based on Strange Thing Mix V4, Strange Thing Mix Anime V1.5, Drochmachina I by Drocheslav

V4 merge based on Strange Thing Mix V3, XAVIER v1.0 by r12x5s and additional LoRAs.

V3 merge based on Strange Thing Mix V2, Strange Thing Variation Style C and additional LoRAs.
V2 merge based on Strange Thing Mix V1, Strange Thing Variation Style A and additional LoRAs.
V1 based on great resources by janxd, n_Arno, Chumple and additional LoRAs.

Search AI Models & Addons

very few results