Live at 3pm GMT 18/08/24

For the comparison of various training settings with the CivitAI website Lora Trainer

FLUX1 DEV. (click show more)

Same dataset with previous models:
SD2.1 https://civarchive.com/models/3727
SDXL https://civarchive.com/models/339881
SDCascade https://civarchive.com/models/320332

Flux Dev Loras:

V1

AssassinKahb-Flux-512x-adamw8bit-e5.safetensors

• Trained with the Civit Flux defaults (17/08/24)

V2

AssassinKahb-Flux-1024x-adamw8bit-e5.safetensors

• Changed resolution to 1024 only, Civit Flux defaults (17/08/24)

V3

AssassinKahb-flux-1024x-Kappa-Prodigy-e12.safetensors

• Using Kappa Neuro's Prodigy Config (see article)

Article: Training FLUX Lora with Civit
https://civarchive.com/articles/6748

Ground Truth

dataset used: https://github.com/MushroomFleet/assassinKahb-1024/tree/main

dataset caption:

AssassinKahb style a demonic looking skeleton holding a sword with red hair 001

dataset example: