Credit:
Trained on this awesome dataset (safe for work) : https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/tsutomu_nihei_lora_training_data.zip
(Created by Adcom: https://tensor.art/u/754389913230900026 )
His Chroma Version of this LoRa: https://tensor.art/models/914986576591739392/Tsutomo-Nihei-e81
----------------------------------------------------------------------------------------------
Trained on NoobAi-XL v1.0 VPRED
tags that have a strong effect:
gritty, gritty background, industrial pipe, swirling, futuristic, scifi themed, ruins,
tag in negative prompt:
watercolor \(medium\)
Version A:
Has received longest training, recommended weight 0.5-08, prompt adherence may suffer from 0.7+ but has a cool visual design. More for dark poster-like aesthetics with swirly forms.
Version B:
Less training, works great at 0.8. Texture is rougher and grittier and colors are brighte than on version A. More detailed backgrounds, better for most normal applications.
Tip: Works very well in combination with the claymore lora (cla 20 vpred version)