Sign In

Loading Images

Well, this image was kind of an experiment, to test how much Flux LoRA I can load at once (more details below).

The image was generated using Flux control net (HED) by X Labs, with settings: control net weight 0.86, guidance 3.5, steps 20. Regarding LoRA, all seven of them were used with varying weights: Bob Ross Art Style with CLIP weight 0,6 and model weight 0,5, DnD Darkest Fantasy — 0.6 and 0.5, Don M Firey Abyss with 0.4 and 0.4, Flux Dark Fantasy 0.3 and 0.3, SXZ Dark Fantasy — 0.3 and 0.4, Flux Realism LoRA 0.4 and 0.4, SXZ Baldur's Gate 3–0.5 and 0.5 respectively. I used ultimate SD Upscale to make 2x upscale with 40 steps, CFG scale 8.0 and denoise factor 0.2. I also cropped the resulted image in an external editor to make a 2048×1536 image.

More technical details on Comfy UI nodes used.

Comfy UI with following nodes:

  • VLM node: Miaoshouai_Tagger

  • Sampler: X labs Sampler

  • Control net: X Labs Control net nodes

  • HED node: HED soft edges

  • Upscale: Ultimate SD Upscale

Models/LoRA/checkpoints used (with respective links to them):

Thanks to the developers of mentioned models and Comfy UI nodes, for inspiration in prompting and workflows. All credits for used models, LoRA and workflows left for the respective authors (AlexLai, Afroman4peace, Artikuz_Ai, ArsMachina, DonMischo, Jedas, sadxzero, Maxfield, adityaintg). Thanks to authors of other awesome nodes, models and tools not mentioned here, but which were essential to create this image.

Comments