<h1 id="ilnai-fujiyama-arashi-from-tokimeki-memorial-girl's-side-3!">[IL/NAI] Fujiyama Arashi from Tokimeki Memorial Girl's Side 3!</h1><p></p><h2 id="v2-noob-ai-(eps):">V2 Noob AI (EPS):</h2><p>Did a re-train</p><h2 id="triggers:">Triggers:</h2><pre><code>f4r4sh1,blonde hair,short hair,spiked hair,brown eyes,</code></pre><h2 id="auxiliary-tags:">Auxiliary Tags:</h2><p><code>anime screencap</code> - Tagged official art work with this! Will skew images to that style!</p><p></p><p></p><h2 id="v1-noob-ai-(v-pred):">V1 Noob AI (V-Pred):</h2><p>Experimental version trained on NAI V-Pred.</p><p></p><h2 id="v1-il:">V1 IL:</h2><p>This was an experiment done for the sake of learning how to train LoRAs. I wanted to limit test how "low effort" I can be with the dataset creation and tagging - basically I skipped a lot of the steps I usually do like image sanitation and in-depth tagging. Aside from a couple of tags, all I did was set a tag blacklist and run the auto-tagger on the site trainer. I had low expectations for this one but it came out surprisingly decent? It performs really well even when just using highres upscaling!<br /><br />Anyway! <br /><br />This is Fujiyama Arashi! He is a judo-obsessed hunk from a Japanese otome game! He is blunt and passionate. He kinda gives me himbo vibes which is why I like him a lot!</p><p></p><h2 id="triggers:">Triggers:</h2><pre><code>f4r4sh1,brown eyes,blonde_f4r4sh1_hair,</code></pre><h2 id="auxiliary-tags:">Auxiliary Tags:</h2><p><code>anime coloring</code> - I tagged high quality CGs with this! Skews images to be more inline with how the game style looks like!</p><p><code>brown_f4r4sh1_eyes</code> - I tagged his eyes with this but I found that it added artifacts sometimes so I usually opt to use only <code>brown eyes</code> instead.</p><p><code>f4r4sh1_uniform,blue jacket,white collared shirt,open clothes,white t-shirt,</code> - Tags for replicating his school look!</p><p></p><h2 id="notes:">Notes:</h2><ul><li><p>This is a LoRA made for my own enjoyment so quality may be variable!</p></li><li><p>Adding facial hair seems to be quite hard. You might need to lower LoRA weight and add <code>mature male, facial hair</code></p></li><li><p>Only intended for Illustrious/Noob AI models!</p></li><li><p>Please use Adetailer to fix artifacts in eyes, hands, etc. 🙇</p></li></ul><p></p>

Fujiyama Arashi (Tokimeki Memorial GS 3)

<ul><li><p>Increased training dataset and step count to 1500</p></li><li><p>Trained on NoobAI XL V-Pred 1.0</p></li></ul>

Files with SHA256: 3967d2bfebb4b35e35914a2aca25d9cbc692fec2f309d0c1814d44bc35944f8b