A LORA in the style of Buchou Chinke, a doujin artist active from the late 90s to the early 2010s, after 2015, Buchou retired from doujin making and drawing but returned to drawing art in 2022. <s>(What Barghest does to a mf)</s>This model specializes in muscular women, if you attempt anything else don't expect the LORA to stay on model, because it won't.Illustrious 2.0 and Pony 2.0 have a completely new, retrained dataset different from previous version with higher steps and a dataset of nearly 300 images, over 100 new images from before with better tagging.NoobAI V-Prediction is out, now! V-Pred is still experimental so expect things not looking right. I recommend checkpoints with Cyberfix merged in to help with anatomy. Also for the V-Pred release I screwed up and set the Epochs too high, so you should probably add monochrome and greyscale to the negatives. Sorry!

[Artist Style] Buchou Chinke

V-Pred version added.Set Rescale CFG to 0.5

buchouchinke-VPRED.safetensors

NoobAI v1.0 V-Prediction

Pony v2

Illustrious v2

NoobAI

v1.0

Beta

A LORA in the style of Buchou Chinke, a doujin artist active from the late 90s to the early 2010s, after 2015, Buchou retired from doujin making and drawing but returned to drawing art in 2022. (What Barghest does to a mf)

This model specializes in muscular women, if you attempt anything else don't expect the LORA to stay on model, because it won't.

Illustrious 2.0 and Pony 2.0 have a completely new, retrained dataset different from previous version with higher steps and a dataset of nearly 300 images, over 100 new images from before with better tagging.

NoobAI V-Prediction is out, now! V-Pred is still experimental so expect things not looking right. I recommend checkpoints with Cyberfix merged in to help with anatomy. Also for the V-Pred release I screwed up and set the Epochs too high, so you should probably add monochrome and greyscale to the negatives. Sorry!

Files with SHA256: 15036e20b5c276b2cdd5dfbd601ca5a95ef32f5f27d1e5634cfa4b014c05c859