A LoRA to improve the appearance of , though you may like it for the tendency to produce smaller . It's not perfect, but it offers an improvement. Sometimes "perfect " helps improve the output.
Version 2 offers:
Improved clarity and definition
In closeups, wrinkles may appear
Improved lighting
More natural skin tones
Improvement in definition.
Note: this LoRA is not intended for gross or graphic depiction of . CivitAI has plenty of LoRAs for that purpose. My overall intention is to produce better and appearance than the models supply.
Description
Anecdotally, Z-Image Base rarely understands or reliably implement "flat-" prompts. The model tends to produce consistent, average, or larger-proportioned, feminine, East Asian-style figures, often ignoring, or struggling to render, prompts for thinner, slighter, or flat-chested body types. It might help to include negative prompts such as , fat-, chesty. Experiment.
Details
Files
Available On (8 platforms)
Same model published on other platforms. May have additional downloads or version variants.