Type | |
Stats | 1,204 18,008 2.5k |
Reviews | (149) |
Published | May 30, 2024 |
Base Model | |
Hash | AutoV2 1061A705A2 |
No triggers needed, suggested weight range: -2 ~ 2 (positive weight means colder, negative weight means warmer) but you can go beyond this range if necessary.
Although it's trained on PonyXL, based on some quick tests, it can also work smoothly in other SDXL models suchs as AnimagineXL or AAMXL.
It may be less noticable in base models with strong coloring such as tPonynai3. It may be due to the training manner for this LoRA since I only trained up layers to make less impact on image composition, but I'm not sure if this is the major cause nor the lbw training is necessary.
Motivation: I noticed that my previous holding camera LoRA looks too warm than usual ones. This is a patch LoRA to fix the issue. It can also work as color temperature slider when used standalone.