<h1 id="sd1.5-82l6by3bu">SD1.5</h1>Previously I was frustrated that none of the model's could produce girl wrapped in towel or a girl wearing a towel. So i tried to find some LoRa but still it was not matching my vision so after countless trial and error i created this LoRa. This produces quite the accurate representation of a girl wrapped in towel. Though there are a few LoRa which produces similar images but still there are not a lot of them so you can try my LoRa and leave a like and comment. It's trained on SD 1.5 you can use other checkpoint and use the one which suits you best<h2 id="setting's-i-used:-04xvwycz6">Setting's I used:</h2><ul><li>Checkpoint: Realistic Vision v6.0B1</li><li>Trigger Word: 1 girl in towel</li><li>steps: 30</li><li>cfg: 5.5</li><li>sampler: euler/karras</li><li>resolution: 512x512</li><li>keep the LoRa strength at 1.0</li></ul><h3 id="note:-e0cj38wc7">Note:</h3>Few color's of towel are not produced properly such as the red color.<h1 id="sdxl-id35i93di">SDXL</h1>All the previous problems with colors are fixed it can produce most of the basic color, it can also produce patterns and multicolor towels. The towel's are much better than the previous version. It's trained on base SDXl model so most of the SDXL based model can work with this LoRa. I don't know if it will work with anime checkpoint's you can try it. If it work's post in the LoRa page.All the setting's are mentioned In the XL LoRa model info tab.<h2 id="trigger-word:-a-girl-in-towel-7a40i0nhf">Trigger word: A girl in towel</h2><h3 id="share-your-creation's-a16dc1wzh">Share your Creation's</h3>If you like my model Leave a comment and like and if you can then buzz me

Towel_Concept_Realistic

I've updated the lora to SDXL the result is much better.I've tested it extensively and it works really well faces might get a bit deformed(Possibility) that's why i created a comfy-Ui workflow that just takes input from the base model and uses a positive text to fix the faces to the already processed image.Hope you have a great time with this new LoRa. Also if you want to request a LoRa you can comment.Settings Recommended<ul><li>Steps" 30</li><li>CFG: 5-6</li><li>Sampler: Euler/Karas or anyone you like try different sampler's to suit your taste</li><li>Resolution: 1024x1024</li></ul>

Files with SHA256: 5b18a6b78cbc124c0d039dfa7f0a8f990b7904b1c21d305340dbdb915eecdcca