BigLove is a merge of TamePony, bigASP and Lustify. It fully understands Pony tags, but interprets them differently. To achieve the greater realism, the Pony weights were reduced in favor of photorealistic ones. It is not meant to do the semi-realism of most realistic Pony models. Still it is better at anatomy then most photorealistic models, but not as good as purer Pony models. The look of bigASP & Lustify is also not fully there, only their photorealistic style.
Big Love is a hybrid. So it produces different images than both Pony & bigASP based models. 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.
For more prompts see the other Big Love model page!
Optimal Settings: DPM++ SDE, Normal / SGM Unifom, 20-30 steps, cfg 3-10 For faster generation 2M SDE/3M SDE also works, but lower quality. For Highest Quality: Use the SDXL Lightning 8-step lora with DPM++ SDE, SGM Uniform, 8 steps, cfg 1.5-2.5
Upscaling: txt2img (Lightning 8-step lora): Hires fix, 4x_NMKD-Superscale-SP_178000_G upscaler, 3 hires steps, 0.4 denoise, 1.5x, hires cfg 2 img2img: same positive prompt, Lightning 8-step lora, DPM++ SDE, SGM Uniform, 1.5x, 6-8 steps, cfg 1, 0.4/0.5 denoise Please note: Unlike txt2img with Hires fix, img2img offer a seed, so you get multiple image versions to choose from. img2img upscaling also produces better quality.
Strengths: * Probably the most photorealistic Pony model * Very good anatomy (See pole dance images) * Huge face variety for a Pony model * Very short prompts are quite effective & photorealistic
Weaknesses: * Still a bit of the Pony look * Depending on prompt it can switch between fully photorealistic and semi-realistic * Sometimes refuses to produce medium shots, portraits or close-ups unless you weigh the term heavily or remove feet, shoes, boots etc. from the prompt. * Sometimes refuses to do **** despite prompting it. Try simplifying the prompt then.
Pony prompts & prompts from bigASP & Lustify work with it. Score tags have as good as no effect usually. A negative prompt is not really needed unless you want to make something vanish or avoid underage. Refer to tamePony and bigASP for further prompting instructions. Pure booru tags and some other samplers will produce Pony style. Photographic terms or a very small prompt ensure a photo look.
The above images are raw output. No img2img, no upscaling, no detailer. You can highly improve the images generated with this model by using these techniques. The Lightning lora enhances the look and sharpness despite only 8 steps. My posts below are usually img2img, so they contain the original prompt, but the settings and seed are different. So you can still reuse the prompt, but you need to change the settings as mentioned above.
Many thanks to the creators of Pony, TamePony, bigASP & Lustify for their wondeful models.
Join https://discord.gg/JCRnbYZm for discussions.
