Sign In

frowning

68
628
518
5
Verified:
SafeTensor
Type
LoRA
Stats
440
470
Reviews
Published
Jun 11, 2023
Base Model
SD 1.5
Training
Steps: 100
Epochs: 2
Hash
AutoV2
D4B5038980
default creator card background decoration
SDXL Training Contest Participant
likot's Avatar
likot

if you wanted to get a slight gradient on the background of the portrait and a portrait stylized as a picture, with photographic blur and juicy lighting, like me, with one click. Then this Lora is for you. Use 512x768 or 912 vertical, training was at these resolutions. increase Lora strength to 1.4 like me or find your own value.

  • use 1.0 to lightly influence the light on the checkpoint.

  • use 1.1 for full CG

Forget about using realistic checkpoint. When you understand that it is impossible to repeat, just download my model SAV_MIX