Sign In

frozen - anna

135
1.2k
24
Verified:
SafeTensor
Type
LoRA
Stats
652
Reviews
Published
Dec 8, 2023
Base Model
SD 1.5
Usage Tips
Strength: 1
Hash
AutoV2
C0A1357185

v2--------------------

Still learning, So update a few of my old Loras with some new training stills :)

Adding more images to the training

Balanced the weight to 1, much less overfit + keeping the face.

Should be much easier to move her around

Do NoT use it with prompts like real, realism, photo real, or something similar, it will make the face less cartoon and it will only damage the result.

You can also add prompts like: "cartoon" "3d" or words like those if you can't get her close enough.

Give it a try :)

hope you like it :D

Join the discord if you like :3

v1--------------------

back to making the Lora version, Loha is too hard to use + too hard to control, unless I need to push some super hard character I don't think I will touch it again :'(

Still learning a lot from making LOHA, This Lora should be able to go higher in weight without going blurry, able to recreate the face just fine, but too close up may cause the eyes crazy big.

try your weight to start around 0.5 this one should be able to go higher to 0.9 but going over 1 will still seem to damage the result.