<h2 id="important-notes:-pk0b3u0s1">Important Notes:</h2>This LoRA is created as part of a fan project for research purposes only and is not intended for commercial use. It is based on the TV series called Arcane which are protected by copyright. Users utilize the model at their own risk. Users are obligated to comply with copyright laws and applicable regulations. The model has been developed for non-commercial purposes, and it is not my intention to infringe on any copyright. I assume no responsibility for any damages or legal consequences arising from the use of the model.<h2 id="compatibility:-266h1im2g">Compatibility:</h2><ul><li>HunyuanVideo</li></ul><h2 id="id-token-trigger-word(s):-2atdzksao">ID Token / Trigger word(s):</h2>Use these in your prompt helps providing the character. See example prompt above.<ul><li>csetiarcane, nfjinx, blue hair, black top</li></ul>Please consider the following:<ul><li>If it doesn't produce the character, try increasing the lora strength to 1.2</li><li>With some seeds it doesn't work well, it simply doesn't produce either the style or the character</li><li>I definitely recommend using at least the following trigger words in your prompt: 'csetiarcane, nfjinx, blue hair' .This is probably due to how the dataset was assembled, I'd like to fix this in later versions"</li></ul><h2 id="acknowledgment:-brwpwdpqc">Acknowledgment:</h2><ul><li><a target="_blank" rel="ugc" href="https://github.com/Tencent/HunyuanVideo">Thanks to the Tencent team for making this great model available</a></li><li>Thanks to <a target="_blank" rel="ugc" href="https://github.com/tdrussell">tdrussel</a> for the <a target="_blank" rel="ugc" href="https://github.com/tdrussell/diffusion-pipe">diffusion-pipe</a> that helps us making these LoRAs.</li><li>Thanks to <a target="_blank" rel="ugc" href="https://github.com/kijai">Kijai</a> for his great <a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-HunyuanVideoWrapper">ComfyUI integration</a></li><li>Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/peteromallet">POM</a> for providing the computing resources. Without this, these LoRAs could not have been created.</li></ul><h2 id="trainig-details:-fc5043mpj">Trainig details:</h2><ul><li>LR: 2e-5</li><li>Optimizer: adamw</li><li>steps: 6000</li><li>dataset: 40 (33x704x352) videos</li><li>rank: 32</li><li>batch size: 1</li><li>gradient accumulation steps: 4</li></ul>

Files with SHA256: 6918f99e75af2629860715d86266ec79ccaa9347e6a2a89662d121d9ce572c78