MOSS

I think I'm getting better at training Moss. Showcase images :

• SFT model + turbo lora 4 steps

• Euler simple / 5 steps / CFG 1

• 1024x1536

• Lora strength : 1

Krea 2

Showcase images :

• Turbo model : 8 steps / CFG 1 & Euler simple

• 768x1152

• x1.5 Hires (10 steps, 0.2 with 4x_NMKD_Siax)

• Lora strength 1 - 1.3

Can also do characters : Tintin, Snowy, Captain Haddock, Professor Calculus.

Ideogram 4

File is big because I tried a lokr 4 to do style + characters, but it failed for the characters (except for Tintin who was already well known in the model).

Example JSON

Anima

I'm having trouble to find good settings for a style lora on this model. For me the quality of this v1 is really meh, but I've already spent too much buzz trying different things (I train everything online). If I try again, I will look for feedback from other trainers on reddit/discord before.

Showcase images :

• er_sde / simple / 10 steps / CFG 1 with turbo Lora

• 768x1152

• Lora strength of 1

Klein 9B

Showcase images :

• Distilled model

• 768x1152

• Euler / normal / 4 steps / CFG 1

• Lora strength of 0.8-1

Qwen:

v2: Retrained on a completely new and higher quality dataset (same one used for training the ZIT version)

ZIT:

Showcase images settings :

• Lora strength 0.7 - 0.8

• DDIM & Beta

• 10 steps & 1 CFG

• 1056x1584

• 'comic style image'

• Long Markdown type prompt for some of them

Flux:

Trained on 30 images handpicked from the comic The adventures of Tintin

It's a style Lora, so it doesnt really know the characters (except maybe Tintin and his dog which Flux already knew a bit).

a bande dessinée style image of

with a weight around 0.8-1. (for Qwen it's 1)