Stable Diffusion 06 Lora Models: Find, Install and Use

Subscribers:
18,400
Published on ● Video Link: https://www.youtube.com/watch?v=o95JkUCbEJ8



Duration: 9:33
7,519 views
121


LoRa stands for Low Rank Adaptation and it is a way to steer the images generated by Stable Diffusion into a specific direction. LoRa's are interesting because of their file size ... where checkpoints are anything between 2-7GB, on average a LoRa is just 100MB. A LoRa can not be used on its own, a checkpoint is always needed.
The prompt used in the video:
(masterpiece, top quality, best quality, official art, beautiful and aesthetic:1.2), 1girl, solo, smile, looking at viewer, long black hair, cns_dress, standing beside a river, (china jiangnan water town style:1.1), beautiful sky, natural lighting
Negative prompt: watermark, text, simple background, white background, EasyNegativeV2
Steps: 24, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 1864753834, Size: 512x512, Model hash: 5493a0ec49, Model: ANI Orangemix3, Denoising strength: 0.32, Clip skip: 2, Hires upscale: 2, Hires upscaler: ANI sudo_RealESRGAN 2x, Lora hashes: "cns_dress_style2_v3: c1cadf46a270", Version: v1.3.2







Tags:
stable diffusion
lora
install
civit ai
model
checkpoint