I made helmetless Doom Slayer or Doomguy LoRA. The reason I made this version because I've tried to make other Doomguy or Doom Slayer LoRA helmetless but only found out that is not what I wanted. Also, I'd like to see the concept of helmetless Doomguy with high accuracy of his face features. I didn't make normal version with helmet because there already existing that such LoRA.Trigger Word ( choose one ):<pre><code>helmetless_doomguy, helmetless_doom_slayer</code></pre>Settings I Used:<ul><li>Sampling Method: DPM++ 2M SDE Karras</li><li>Sampling Stpes: 22 - 25</li><li>Weight: 0.65 - 0.9</li><li>Resolution: 768x768, 960x960, 768x960</li><li>CFG Scale: 7</li></ul>If you like my LoRA, you can buy me some coffee:<a target="_blank" rel="ugc" href="https://www.buymeacoffee.com/doomguy1111">https://www.buymeacoffee.com/doomguy1111</a>

Helmetless Doom Slayer | DOOM Eternal | SD 1.5

I re-checked the model and I found out that this model is extremely rigid and hard to worked on if you want to strip his armor and change to other clothing. Also, I found out there's some bad images in my dataset. So, I decide to delete them and/or change them into higher resolution. Here's some fix that I made:<ul><li>Decreased dataset size from 110+ to 80+<ul><li>I hope this image quantity with higher quality, improves the image quality</li></ul></li><li>Fixing mislabeled or missing label during dataset labelling</li><li>Add support to other clothing. In other word, increased flexibility of the model</li><li>Decreased LoRA size slightly for adjusting dataset's change</li></ul>

Files with SHA256: 83506d65ddc9e0e15a5aaae7267db7536b17b39c39d0579d72f8dd0af3a23c35