Elsewhere XL

At the Facebook Pure Fooocus group, I developed a formalized testing procedure for base models. I have published the results for a wide variety of models, and FaeTastic 2.4 remains at the top of the list, even surpassing the current Flux offerings as a general purpose image creator.

Guided by this test information, I attempted to create a model that was optimized for my particular interests, and this merge is created from SDXL FaeTastic 2.4 (48%), JuggernautXL XI (32%) and animaPencilXL 5.0 (20%) together with the SDXL Offset Example (0.5) and Better Blur Control (1.0) LoRAs. The goal was to create a model that is suitable for both fantasy and realism - or even better, for realistic fantasy or fantastic reality.

I am very grateful for each of these creations. They each represent an enormous amount of dedicated hours and inspired problem solving, shared freely here at CivitAI. Please respect the licenses of each of these creators. In particular, do not use this model on any service that monetizes image creation, such as for online image generation. And also, do not sell or license this model for a fee or something else of value.

Usage:

Please use your preferences for samplers and other parameters. My choices for Elsewhere XL are a CFG of 4.5 (going up to 8 for special effects), sampling steps at 30, and the default Fooocus sampler, dpmpp_2m_sde_gpu karras.

Put "sharp background" in the positive prompt and "blurry background" in the negative prompt to increase the depth of field. "Bokeh" or "blurry background" in the positive prompt creates a shallow depth of field.

I have absolutely no interest in NSFW imagery. If this model creates NSFW images, it is not intentional.

Credits:

SDXL FaeTastic 2.4: https://civitai.com/models/129681

Juggernaut XI by RunDiffusion: https://civitai.com/models/133005?modelVersionId=782002

animaPencilXL 5.0: https://civitai.com/models/261336/animapencil-xl

Better Blur Control LoRA: https://civitai.com/models/324815/better-blur-control-less-depth-of-field-xl-lora

SDXL Offset Example LoRA: https://civitai.com/models/137511/sdxl-offset-example-lora

The many test merges involved in this project were all created with the RuinedFooocus MergeMaker, a simple but effective tool: https://github.com/runew0lf/RuinedFooocus/wiki/Checkpoint-merges

The featured images are samples from the ten standard tests applied to this particular model at the Pure Fooocus Base Model Reviews Guide: https://www.facebook.com/groups/fooocus/learning_content/?filter=519238867322550&post=753347100153801
(Facebook links are notoriously unreliable, you may need to scroll up or down to access this particular guide)