Style LoRA of Nusmusbim / ใในใ (ใในใ ในใใ )
Like the style? Give the artist some support!
โค๏ธ Drop a ๐ & follow me if you want more!
I look forward to your posts, reviews and comments!
๐ Recommendations
Use CFG++ samplers such as Euler CFG++ with a CFG of 1-2.
If no trigger tags are specified, this LoRA does not need a trigger tag to work!
Do note that my LoRAs that have been baked on Illu0.1 are compatible with NoobXL's V-Pred models (and any merges / finetunes).
๐ Training Notes
Moved to the 'About this version' tab!
Description
๐ Training Notes
Base Model: NoobAI-XL V-Pred 1.0
Training Resolution: 1024
Training Batch Size: 4
Type: Lora
DIM / Alpha: 16 / 8
Conv Dim / Alpha: 24 / 12
Train On: Unet
Optimizer: Prodigy
Scheduler: Rex
Image Count: 348
Steps: 1900