In the pony test points I tested, it is very difficult to represent middle-aged people.

This LoRA is an attempt to get around the problem.

The trigger is "milf" and for information such as "40yo", "50yo" or "60yo" the effect is increased. The weight of the LoRa should not exceed 0.7.

Update V2:

new detailed tags, optimized training photos,

The LoRA now offers significantly improved age recognition and more flexible characters.

The LoRa was trained with 462 training photos.

If you like the LoRa, I would be happy if you gave it a "thumbs up".

Please post your results directly here on the model page.