Stable Diffusion

Illustro

model_id: illustro

<h2 id="heading-1284"><span style="color: rgb(253, 126, 20);">NOW OVER </span><u><span style="color: rgb(253, 126, 20);">9,700</span></u><span style="color: rgb(253, 126, 20);"> COMBINED downloads!!!</span>❤️<br><em><span style="color: rgb(250, 82, 82);">per sacrificium purificant</span></em></h2> <p>This model is the next evolution of a pure-photorealistic fantasy model. Capable of creating any character &amp; any setting you can feasibly conceive!<br><br>It's the most detailed &amp; thoroughly made model I've released as of yet &amp; I spent tireless nights fine-tuning this to be the most epic model I could manage. I spent roughly a month generating examples and running it through its paces. It's also capable of using most LORA's quite effectively so you should be able to create whatever you like.<br><br>ALL THE MAIN EXAMPLES ARE 1:1 FOLLOWING MY INSTRUCTIONS YOU SHOULD BE ABLE TO REPLICATE THEM EASILY!</p> <p>&nbsp;</p> <h3 id="heading-434">This model can do pretty much any style fantasy art at a very high level. The main concepts:</h3> <p>&nbsp;</p> <p>✔️ <strong>fantasy art (retro &amp; modern)</strong><br>✔️ <strong>characters / rpg characters</strong><br>✔️ <strong>photorealistic / realistic</strong><br>✔️ <strong>analog style</strong><br>✔️ <strong>surrealism</strong><br>✔️ <strong>oil painting</strong><br>✔️ <strong>concept art</strong><br>✔️ <strong>cinematic</strong><br>✔️ <strong>scifi/science fiction</strong><br>✔️ <strong>landscapes &amp; scenery</strong><br><br></p> <p><em>(all of the above list can &amp; should be used as triggers and weights)</em></p> <p>It knows most common <strong>D&amp;D races</strong> and it will produce some interesting great results when prompted correctly. (It has some minor issues with complicated races such as dragonborn &amp; tieflings, etc. but you should be able to use a LORA to render these effectively. &lt;3)</p> <p>&nbsp;</p> <h2 id="heading-1285"><em><span style="color: rgb(250, 82, 82);">How to use</span></em><br><br><strong><em>There is no embedded VAE but it was designed to be used with this:</em></strong></h2> <p>☞ <a href="https://huggingface.co/stabilityai/sd-vae-ft-mse-original/tree/main" target="_blank" rel="noopener ugc"><strong><u>vae-ft-mse-840000-ema-pruned.ckpt</u></strong></a></p> <p><strong><em>I also recommend using the following negatives:</em></strong></p> <p><a href="https://civitai.com/models/7808/easynegative" target="_blank" rel="noopener ugc"><u>EasyNegative</u></a> or <a href="https://huggingface.co/gsdf/Counterfeit-V3.0/blob/main/embedding/EasyNegativeV2.safetensors" target="_blank" rel="noopener ugc">EasyNegativeV2</a> (The all around catch all. Always good in a bind)<br><a href="https://civitai.com/models/16993/badhandv4-animeillustdiffusion" target="_blank" rel="noopener ugc"><u>badhandv4_anime</u></a> (absolutely your go to neg for anime hands)<br><a href="https://huggingface.co/yesyeahvh/bad-hands-5/tree/main" target="_blank" rel="noopener ugc"><u>bad-hands-5</u></a> (good for the more realistic images but will not work amazingly for anime)<br><a href="https://civitai.com/models/5224?modelVersionId=6057" target="_blank" rel="noopener ugc"><u>bad-artist \ bad-artist-anime</u> </a>(I&rsquo;ve had good and bad luck)<br><a href="https://huggingface.co/datasets/Nerfgun3/bad_prompt/tree/main" target="_blank" rel="noopener ugc"><u>bad_prompt \ bad_prompt_version2</u></a> (works very good sometimes)<br><em>(I typically use (bad-hands-5) &amp; (bandhandv4_anime) at the same time to cover all bases.)</em><br><br><br>(Here are some example settings for you to start with, but please keep in mind YMMV!)</p> <p>&nbsp;</p> <p><strong><em>Steps:</em></strong> 30 = Sweet Spot / 10-150<br><strong><em>Sampler:</em></strong> DPM++ 2M Karras, DPM++ SDE Karras, Euler A, DDIM,<br><strong><em>CFG scale:</em></strong> 4-11.5 = Sweet Spot / 4.5-15<br><strong><em>Size:</em></strong> 512x768, 768x512, 640x840, 840x640, 600x896, 896x600, (any size)<br><strong><em>Clip skip:</em></strong> 1<br><strong><em>Hires Fix:</em> </strong>Enabled<br><strong><em>Upscale by:</em> </strong>1.5-2<br><strong><em>Hires Steps:</em></strong> Equal to Half or Exact the exact amount of Steps (or higher)<br><strong><em>Upscaler:</em> </strong><a href="https://upscale.wiki/wiki/Model_Database#Universal_Models" target="_blank" rel="noopener ugc"><strong><u>Remacri Upscaler</u></strong></a>, <a href="https://nmkd.de/shared/?dir=ESRGAN/Models/Art" target="_blank" rel="noopener ugc"><strong><u>Lollypop Upscaler</u></strong></a>, <a href="https://nmkd.de/shared/?dir=ESRGAN/Models/Art" target="_blank" rel="noopener ugc"><strong><u>4x_NMKD_DigArt_10k</u></strong></a>|<br>( <em>I use </em><strong><em>DigArt 10k </em></strong><em>almost exclusively with everything except for 2D illustration &amp; animation</em> )</p>