pony nsfw on Anyken
pony nsfw on Anyken: compare the model examples with your intended input and output. Review the capabilities documented below, then use the controls actually offered in the workspace. Available settings and credits are shown before submission.
Read source
pony nsfw: Model Overview
Pony Diffusion V6 XL is a versatile SDXL finetune capable of producing stunning SFW and NSFW visuals of various anthro, feral, or humanoids species and their interactions based on simple natural language prompts.
Model Architecture and Training
Pony Diffusion V6 XL is built on the SDXL architecture and has been trained on approximately 2.6 million images, aesthetically ranked based on the author's personal preferences. The training dataset maintains a roughly 1:1 ratio between anime, cartoon, furry, and pony datasets, as well as a 1:1 ratio between safe, questionable, and explicit ratings. About 50% of all images have been captioned with high quality detailed captions, resulting in very strong natural language capabilities. All images have been trained with both captions (when available) and tags, while artists' names have been removed and source data has been filtered based on an Opt-in/Opt-out program. Any explicit content involving underage characters has been filtered out. The model supports a wide array of styles and aesthetics but provides an opinionated default prompt template that allows generation of high quality samples with no negative prompt and otherwise default settings.
Prompt System and Quality Modifiers
The model uses a specific prompt template that begins with quality modifiers: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up. Previous Pony Diffusion models used a simpler score_9 quality modifier, but the longer version in V6 XL is a training issue that was too late to correct during training. Users can still use score_9 alone, but it has a much weaker effect compared to the full string. The model is designed to not need negative prompts in most cases and does not need other quality modifiers like "hd" or "masterpiece". Other special data selection tags include source_pony, source_furry, source_cartoon, and source_anime, as well as ratings of rating_safe, rating_questionable, and rating_explicit. The model is capable of recognizing many popular and obscure characters and series. For pony style specifically, the recommended templates are either "anthro/feral pony, rest of the prompt" or "source_pony, rest of the prompt".