Oldhammer-SDXL-LoRA-Warhammer-40k

<html> <head> <link href="https://fonts.googleapis.com/css2?family=Skranji&display=swap" rel="stylesheet"> </head> <style> .title-container { display: flex; justify-content: center; align-items: center; } .title { font-size: 2.5em; text-align: center; color: #1d1d1d!important; font-family: 'Skranji'; text-transform: uppercase; letter-spacing: 0.1em; padding: 0.5em 0; background: transparent; } .title span { background: -webkit-linear-gradient(45deg, #fcb900, #ff6900); -webkit-background-clip: text; -webkit-text-fill-color: transparent; } .icon { margin-right: 10px; } .icon img { width: 50px; height: auto; padding-bottom: 12px; } </style> <div class="title-container"> <div class="icon"> <img src="https://huggingface.co/Schisim/Oldhammer-SDXL-LoRA/resolve/main/assets/skull-01.svg" atl="Skull Icon"/> </div> <h1 class="title"> <span>Oldhammer SDXL</span> </h1> </div> </html> <Gallery /> <!-- START SUPPORT --> <a href='https://ko-fi.com/M4M3ZBOIT' target='_blank'> <img height='36' style='border:0px;height:36px;' src='https://storage.ko-fi.com/cdn/kofi3.png?v=3' border='0' alt='Buy Me a Coffee at ko-fi.com' /> </a> <!-- END SUPPORT -->

The inference images are just to display samples. The inference API is not connected.

<!-- START Model Info --> <div class="model-info-container" style='font-family:Monospace'> <table id="model-info-table"> <tr> <th></th> <th></th> </tr> <tr> <td>Base Model</td> <td>SDXL v1.0</td> </tr> <tr> <td>VAE</td> <td><a href="https://huggingface.co/madebyollin/sdxl-vae-fp16-fix">SDXL-VAE-FP16-FIX</a></td> </tr> <tr> <td>Type</td> <td>LoRA</td> </tr> </table> </div> <!-- END Model Info -->

Note:<br> This LoRA was trained on a subset of a much larger dataset I am creating for a full-parameter WH40k SDXL fine-tune.<br> The dataset was not designed for LoRA training. As a result, you may experience some 'wonkiness'.<br> The prompts for the sample images are mostly taken from the dataset caption-pairs (the subset has been recaptioned since this LoRA was created). You can get much better results using your own prompts.