Aging Slider for ANIMA:
Positive weights (+1.0 and above): Realistic aging. Wrinkles, sagging skin, wisdom — the full grandma experience.
Negative weights (-1.0 and below): Youth mode enabled. Pushing this into negatives will effectively de-age the character.
⚠️ DISCLAIMER / USAGE POLICY: Please note that the negative side of this slider is NOT intended for creating NSFW content. This is simply a technical age regulator. Please strictly adhere to the platform's rules and Terms of Service when using this tool. Use it responsibly and keep it clean.
Support & Exclusive Content
You can find my guide to creating LoRa Styles, Early access and Exclusive models on my https://boosty.to/blacksnowskill or just support me 💜
https://t.me/blacksnowskill - news
License & Usage
Notice: This model is protected by copyright. Copying, distribution, merging with other models, or use on other websites without the explicit written permission of the author is prohibited.
Description
FAQ
Comments (3)
Based on the example, this LoRA seems to drastically alter the image style. Negative values appear to push the style increasingly toward "anime," while positive values seem to shift it toward "Western comics" or a semi-realistic look. Is that actually the case? It could quickly become an issue regarding the consistency of the generated images.
Thank you for the detailed feedback and for noticing this behavior.
This is indeed a characteristic of the Anima base model during LoRA training. Since the training is done purely on vector differences (positive = older, negative = younger) without using reference images, the model sometimes amplifies certain stylistic tendencies: negative values tend to push toward a cleaner anime aesthetic and lighter tones, while positive values shift toward a more Western comic / semi-realistic look with darker tones.
This is the most stable and effective age slider I've trained after 10+ attempts, but I'm actively working on taming these side effects to improve style consistency across the range.
Appreciate you pointing it out — feedback like this helps a lot.
This is expected, because real photos of underage subjects are nearly absent in the training of the base model. In other words, the model "thinks" a child is a cartoon, whereas for adult subjects the source images during training are mixed. This and most age sliders are more useful in the adult direction. Which might seem limiting but at the end of the day is good because it helps keeping you out of legal trouble, most of the time.
