No special tags needed, just prompt normally while the LoRA is enabled. "Hires fix" is recommended for best results.
If you use this LoRA as part of a mix or host it on a generation service, please mention and link back to this page (especially if you're making money off of it)
Description
460 images
Tagged with wd14-swinv2-v2
Tagged image repeats:
111 1x repeat "close-up"
349 2x repeat "masterpiece, best quality"
x768 base resolution with aspect ratio bucketing
20 epochs, batch size 4
1e-3 cosine with restarts LR (matching U-Net and text encoder)
20 cycles
256 dim
Clip level 2
[redacted] base model
[redacted] .vae enabled during training