<h1 id="heading-11">This model has been relocated to 🤗 Hugging Face.</h1><h2 id="heading-12"><a target="_blank" rel="ugc" href="https://huggingface.co/JerryOrbachJr/Franklin-Booth-Style">https://huggingface.co/JerryOrbachJr/Franklin-Booth-Style</a></h2><hr /><h2 id="heading-201">About the Artist</h2>Franklin Booth (July 8, 1874 – August 25, 1948) was an American artist known for his detailed pen-and-ink illustrations. His unique style was formed by his early practice of copying wood engraving illustrations. His skill as a draftsman and recognizable style made him a popular magazine illustrator in the early 20th-century.<h2 id="heading-202">The LoRA</h2>This LoRA was created using 40 high-resolution scans of some of Booth's best work and can add interesting line-shading effects and other aspects of Booth's style to your images. It's a work in progress, and feedback, including suggestions, is welcome.<h2 id="heading-1142">Settings</h2><ul><li>Dimensions: Bigger images will generally require lower strength - at lower resolutions it seems like SD is combining lines into a gray blob. 512 in either dimension should be a minimum, and if you can do 768+ (with or without "Kohya Shrink Wrap"), you will likely get better results.</li><li>CFG: the lower the CFG, the less strength you will need to see the effect. I suggest starting at 3.5 and going down from there if possible. 6 is probably the highest I've used with this.</li><li>Fighting/Working With the Style: the more old-timey, pen-and-inky, and realistic your prompt is, the lower the strength you will need. Concepts with a lot of round shapes will need a higher strength. If you add "black and white engraving" to your prompt it's usually like adding +0.5 to the strength without losing image quality, so give that a try if the strength gets to high and your image is suffering</li><li>Strength: taking all that into effect, you will usually need a weight of between 0.5 and 1.5 to get a good effect with this LoRA. Start at 1 and see how it goes!</li></ul><h2 id="heading-630">Compatible Models</h2>This LoRA should work with a very wide variety of models, although it will likely not work well with models trained exclusively on amine (if there are models out there trained on manga I'd bet it would work well with them, though). It's been tested and has created some good images with the following models:<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/132632?modelVersionId=223670">epiCPhotoGasm - Last Unicorn</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/84728?modelVersionId=90072">Photon v1</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/4384?modelVersionId=128713">DreamShaper v8</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/15003?modelVersionId=256915">CyberRealistic V4.1 "Back to Basics"</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/97744?modelVersionId=252613">RealCartoon-Realistic v12</a></li></ul>

Franklin Booth Style

v1 - Poorly trained If your prompt includes words like "boat" you will get a very specific kind of boat (the ones in the training images). And if you ask for things far away from the training images, you really have to crank up weight and play with cfg and stuff, if it will work at all. Settings: Include "black and white engraving" in your prompt. Try starting with 1.5 weight and CFG 3.5 with at least 30 steps.

Files with SHA256: 29351f5205df7018de5415a7f07528cb0627a3c58c74d02d01516864b1c56e11