<h2 id="heading-2781">トーラムオンライン/Toram Online</h2><h3 id="heading-2782">A video game graphics style LoCon</h3>This LoRA tries to replicate Toram Online MMORPG graphics.<a target="_blank" rel="ugc" href="https://toram.jp/">https://toram.jp/</a><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1d7c0632-b9ee-4c12-b39f-b687c6b212ac/width=525/1d7c0632-b9ee-4c12-b39f-b687c6b212ac.jpeg" />credit image by: @XenGamingChannelTrained with 343 images using various in-game screenshots and images found online. Model created for research purposes, but you can use this for your YT thumbnails if you want to. works better on female characters, Idk why.<h3 id="heading-2783">Recommeneded strength</h3>~0.8 The eyes start to decay below and above this number, idk why.<h3 id="heading-2783">Recommeneded prompts</h3><ul><li>1girl, solo</li><li>1boy, solo</li><li>mix or multiple (ex. 2girls, multiple girls, 3boys, multiple boys)</li><li>looking at viewer</li></ul><h3 id="heading-146">Recommeneded Negative prompts</h3><ul><li>worst quality, lowres</li><li>from side (if it keeps showing side views)</li></ul>You can use the generated images in the gallery as a guide. Prompts will be included.This is the first LoRA that I am willing to publish and I know that it needs a lot of improvements, so please let me know what you guys think. Thank you!!!

トーラムオンライン Toram Online MMORPG | Style

This is a fix for v1.0. What did I changed? I just removed all datasets that contain multiple people. Apparently, even SDXL cannot generate multiple people properly.

Files with SHA256: cb24ea857fb6f40d63d18d4458ebbd5cdecd6a8ccf56f3cbdc518b706e8c66e3