Big Love (XL 2.5)

https://civitai.com/models/897413?modelVersionId=1590591

BigLove is a merge of Pony, bigASP and other SDLX models. The Pony releases of Big Love are a hybrid. They produces different images than both Pony & SDXL-based models. Pony tags are understood, but interprets them a bit differently. Some images lean more in the one or other direction depending on how you prompt. It works with Pony as well as normal SDXL loras. It can turn anime/cartoon into photorealistic images in img2img.

Big Love XL is a combination of 5 different training branches of SDXL: original SDXL, bigASP, RealVis, Anteros and Pony. With a bit of Pony in it, some Pony loras (poses, characters) work too. It outputs a more photorealistic & creative look than the Pony versions and is can also do paintings & cartoon.

txt2img Settings: Forge/ComfyUI: DPM++ SDE, Normal/SGM Unifom (Pony), Karras (XL), 20-30 steps, cfg 3-10, clip skip 2, 832x1216 or 1024x1536. For faster generation 2M SDE/3M SDE works too, but lower quality with the Pony version.

Civitai: DPM2 a, 30-50 steps, cfg 4-5, 832x1216

Higher Quality & Faster Rendering: Good Quality: For more contrast add the SDXL Lightning 8-step lora with DPM++ SDE, SGM Uniform, 8 steps, cfg 1.5-3 (Pony), 1-2.5 (XL), lora strength 1.0, clip skip 2, 832x1216 or 1024x1536 Best Quality: For a more natural look use the DMD2 lora with LCM Exponential, 8 steps, cfg 1-1.5, lora strength 1.0, clip skip 2, 832x1216 or 1024x1536

Upscaling:

1. txt2img with Lightning 8-step lora: Hires fix, 4x_NMKD-Superscale-SP_178000_G upscaler, 3 hires steps, 0.4 denoise, 1.5x, hires cfg 2, clip skip 2 Higher Quality & Faster Rendering: Good Quality: img2img with Lightning 8-step lora: same prompt, DPM++ SDE, SGM Uniform, 1.5x (XL), 2x (Pony), 3 steps (ComfyUI), 6-8 steps (Pony), 8-12 steps (XL), cfg 2, lora strength 1.0, 0.3-0.5 denoise, clip skip 2 Best Quality: img2img with DMD2 lora: same prompt, LCM Exponential, 1.5x, 8-12 steps (XL1/XL2) 4-8 steps (XL2.5), cfg 1, lora strength 1.0, 0.3-0.5 denoise, clip skip 2 Please note: Unlike txt2img with Hires fix, img2img offers a seed, so you get multiple image versions to choose from. img2img upscaling also produces better quality in A1111/Forge.

ComfyUI Workflow: Download the following image and drop it into Comfy: Lightning Workflow DMD2 Workflow (recommended) DMD2 UltraRes Workflow With the Lightning lora it is better to use Forge. ComfyUI is not able to do as many upscaling steps with Lightning as Forge without adding grain. But if you use the DMD2 lora, ComfyUI does just fine. The UltraRes workflow also requires the UltraRes lora for hires output.

Strengths: Great fine details, very natural looking photo-like images, lovely faces, creative poses, can also do paintings and cartoon

Weaknesses: Some prompts produce softer skin, multi-person images can have strange anatomy

Prompting Advice: SDXL prompts & prompts from Acorn Is Boning & Lustify work with it. Pony prompts too, but they create a more photorealistic look. Natural language prompts tend to create artistic glamour images with less skin detail while normal tags do more photo-style images. Big Love XL gives individual words more attention, so you don't need to rework prompts like Big Love Pony requires sometimes. XL2.5 can sometimes get too sharp. Reduce cfg or use less steps in img2img to make images softer.

Using Character Loras: Big Love Pony supports Pony characters, but they look like real humans. There were some complaints about them not working, but the problem was not the model itself so far. So here are some tips:

• Try Pony as well as SDXL character loras
• Use a lora weight up to 1.5 if it does not reduce quality
• If it is a well known character, add the name in the prompt and also give it a higher weight if necessary.
• Check your negative prompt for problems. Delete it if necessary and rewrite it from scratch.
• Render a lot of images to be able to pick those with the best resemblance.

Higher Ouput Quality: You can highly improve the images generated with this model by using img2img, upscaling, detailing etc. The Lightning lora enhances the look and sharpness despite only 8 steps. Alternatively, the DMD2 lora produces a sharp, but more natural look with less contrast. Also check out the Subtle Style loras, which nicely enhance images.