Greetings!Now this is a really special one, mostly because I couldn't find any other LoRa that could replicate GB's style. What I really wanted to do was to replicate his style with any characters, but I'm still learning on how to do that properly (styles are trickier!).<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/9ad9917f-35ab-49c4-80a2-4bdd8a83cf23/width=525/9ad9917f-35ab-49c4-80a2-4bdd8a83cf23.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/f30b6077-2c3c-49c0-9bb1-97d153ab14de/width=525/f30b6077-2c3c-49c0-9bb1-97d153ab14de.jpeg" /><h1 id="heading-1708">Model Characteristics</h1>For now, enjoy this instead. I might also do a version 2.0 on this but for now I'm happy enough with the results. There are a few things to keep in mind though:<ul><li>Trigger word: mercy_\(overwatch\)</li><li>Use weight of 0.7 to 1. See what works best for you.</li><li>I'm using <a target="_blank" rel="ugc" href="https://civarchive.com/models/58431">DarkSun v7b </a>and <a target="_blank" rel="ugc" href="https://civarchive.com/models/94809">RealCartoon3D v13</a> and got good results, so I'd recommend using checkpoints that have a Semi-Realistic style.</li><li>Full body shots appear to mess her face up a little, so I'd recommend using ADetailer.</li><li>Restore face is optional (ADetailer works way better).</li><li>Play around with the samplers, but I got best results with DPM++ 2M Karras and DPM++ SDE Karras.</li><li>Clip Skip 2 works best.</li><li>To avoid overfitting, I recommend using the extension <a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-lora-block-weight">Lora Block Weight </a></li><li>Hires Fix works like a charm, I recommend using R-ESRGAN as the upscaler.</li></ul><h1 id="heading-1709">Trainning</h1>I'm relatively new to LoRa trainning so any tips for improvement are much appreciated!<ul><li>Dataset size: 84</li><li>Model: <a target="_blank" rel="ugc" href="https://civarchive.com/models/4384/dreamshaper">DreamShaper 8</a></li><li>Buckets enabled</li><li>Epoch: 1</li><li>Repeats: 100 (good dataset) and 50 (bad dataset)</li><li>Total steps: 2088</li></ul>I'd be happy if you guys could give me any character suggestions for my next LoRA!Cheers.Like my stuff? Buy me a Monster!<a target="_blank" rel="ugc" href="https://ko-fi.com/lizarb"><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1ae16ea7-0a66-4765-a795-b9681a98a4ae/width=525/1ae16ea7-0a66-4765-a795-b9681a98a4ae.jpeg" /></a>All my models and images will always be 100% free, don't feel obligated to donate!

Mercy (Overwatch) Generalbutch style - LoRa

<ul><li>Improved dataset</li><li>New training settings should now make the model more flexible</li><li>Overfitting was slightly reduced (works better with extension <a rel="ugc" href="https://github.com/hako-mikan/sd-webui-lora-block-weight">Lora Block Weight</a>)</li></ul>

Files with SHA256: 55f28d95c231903da238e9737f35f54000f6ca225998e66584ec1da6a22dbd8d