I'll try to update again in June, as I ran out of Colab credits and wanted to take some time off.
The current model is in the experimental stage (I am learning how to train the Lora model).
Probably v0.3.0 is the best version. This needs to be chosen according to your personal preference.
If you think this Lora model is not very bad, then thank you very much.
If you like this Lora model, you may consider clicking the like button.
I would recommend mixing multiple Lora versions. At the moment I can not make a very good Lora model.
class_tokens = "makise kurisu 1girl"
If you want to generate details such as makise kurisu outfit and pupil color, I prefer to use: ultra detailed, 1girl, makise kurisu, slim, tall, auburn hair, soft curls, green eyes, labcoat, red tie, professional, serious, black shorts, black socks, brown loafers, sleek, sophisticated, stylish, <lora:makise kurisu:0.6>,
Description
The v0.4.0 version uses a mask to divide the foreground and background, which I think has some effect; when you set too high a weight, like <lora:makise kurisu:0.8>, you'll see a translucent mask between the hair gaps.Unless you set a suitable weight, such as <lora:makise kurisu:0.6>.
I'm sorry if you see the hand details are terrible.I don't know how to fix the detailed performance of the extremities.I don't want a lot of training steps at the moment to address the concept of fingers and toes.If you have prepared some means of fixing it yourself, that's great.
FAQ
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.