This is the Illustrious version of my <a target="_blank" rel="ugc" href="https://civarchive.com/models/1059949?modelVersionId=1189481">Pony Rough Sketch Style. </a>It's supposed to work the same way but it has a more "finished" look. There are 3 versions of this model as I could not decided between 3 different epochs: V1 - Epoch 1, V2 - Epoch 7, and V3 - Epoch 10. I'll let you choose.I'm using reForge to generate the images and I'm a noob in training LoRas in general.This Lora was trained in regular Khoya SS using this script: <a target="_blank" rel="ugc" href="https://pastebin.com/ZYHjafvU">https://pastebin.com/ZYHjafvU</a> I found this script from this article on the comments section: <a target="_blank" rel="ugc" href="https://civarchive.com/articles/9148/illustrious-lora-training-discussion">https://civarchive.com/articles/9148/illustrious-lora-training-discussion</a>I used the original Illustrious XL model for training as I tried using NoobAi XL vPred 1.0 on site and it didn't work as I hoped for. Anyway these are the information about the Lora.Trigger Word: R0u6hSk3tch, sketch, greyscaleNegative Prompt: lowres, (worst quality,bad quality,low quality:1.2)Sampler: Euler and/or DDPMScheduler: NormalWeight: 1Model Used: NoobAI-XL (NAI-XL) v-pred v1.0Do not use tags such as "bad anatomy" or "bad hands" or "simple background" or anything similar in negatives as the point of this model is to have an unfinished look while keeping the good quality of the generation.I used samplers Euler and DDPM to test this model and me personally liked the generations with both. Some prompts looked better in Euler and some prompts look better in DDPM.Make sure to post your favorite generations on the Gallery by simply clicking +Add Post button.Until next time.

Files with SHA256: 7dc66ffbbc7e65862bffd5b70696e14b2f763268b99bdc2f810757a66b474640