<p><strong><em>V2</em></strong></p><p>This time the LORA was trained on WAI (v12.0) as target model. I cleaned the dataset a bit to have better training data, which resulted in a style more close to the OG videos. To avoid the unwanted brains showing randomly in the generated images, the trigger word is now "kanaria_brain_style".</p><p></p><p><strong><em>V1</em></strong></p><p>Style based on Kanaria's Brain MV. I extracted equidistant frames from the videos at the bottom.</p><p>The trigger word is "brain-style". Because of the "brain" in the trigger, this lora tries to add brains to the image; I plan on training a V2 with a different trigger (when I save enough buzz). In the meantime, try adding "brain" and "brain exposed", as negative prompt. Also, I trained on vanilla Illustrious; I works fin with JANKU, and I plan to train V2 with WAI-NSFW as I usually do. <br /><br /></p><div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/h4HkXR3NSI4" start="0"></iframe></div><div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/rRNjcKMhlu8" start="0"></iframe></div><div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/LplOoRhetlc" start="0"></iframe></div><p></p><p></p><p></p>

Files with SHA256: bc328e692c3b558f569649946a6b26e2493f79537800d4032b2458daa7e70db8