Sign In

1930s Animation Style XL | Distrill

102
667
1.2k
19
Verified:
SafeTensor
Type
LoRA
Stats
667
1,226
205
Reviews
Published
Sep 7, 2023
Base Model
SDXL 1.0
Training
Epochs: 25
Hash
AutoV2
C241B4A68B
default creator card background decoration
distril's Avatar
distril

1930s Animation Style XL

Consider supporting further research on Ko-Fi

No activation word required

Tips:

  • Works best with a CFG of 8-12

  • Use LoRA weight of around 0.7 as this version is a little crispy for my taste.

  • I recommend DPM++ 2M and DPM++ SDE at 28 steps

  • Higher CFG will result in more solid colors/cartoony, lower CFG is more life like and can get weird.

  • No style prompt required

  • No trigger keyword require

Information:

This was trained with 32 images using an adaptive optimizer and only the Unet was trained for this version. The images was from various sources that represented the iconic style of the Golden Age of American Animation. Examples are Max Fleischer, Walt Disney, Felix the Cat, Cuphead, Old school mario art (nintendo modeled him with the style in mind).

I personally like the look of using a LoRA weight of between 0.4-0.8 and a CFG of 8-12 since I prefer more 2D cartoony looks. If you want a weird mix of real life and cartoon you can get that with a CFG of 6-8 but lower the LoRA weight if you use lower CFG as it gets messy otherwise.

License
The license type is CC BY-NC-ND 4.0. If you want to use the LoRA on your service and/or for commercial purposes, message me on Discord. The philosophy is very simple – private individuals provide support for the model, not companies, so private individuals are given liberties.