Discover MY ABOUT & LINKS to find more my models and art.

It is one of my first LoRAs. It is very experimental and probably far not perfect.

I used dataset of 36 images 768x768 resolution. I put in dataset only photos of Kristen Stewart with brunette hair. I also have Kristen Stewart blonde version.

Trigger word: AIDA_LoRA_KristenBrunette

Recommended values: (AIDA_LoRA_KristenBrunette:1.33) <AIDA_LoRA_KristenBlonde:0.6>

LoRA trained using kohya_ss on 'v1-5-pruned-emaonly' checkpoint, so it should works well with most of v1.5 forks and mixes. Training steps: 1512. Learning rate: 0.00001. Precision: fp16. LR Scheduler: constant_with_warmup. LR warmup: 30%.

Purchasing "LoRA of Kristen S. (brunette version) (For SD1.5)." in MY SHOP for 1$ you can say me "thank you" for this model and achieve rights for selling images generated with it.

Please contact me via [email protected] or via X DM if you have any claims/questions or if you need some extended rights for sharing, selling, merging this LoRA and so on.