<h2 id="please-give-a-thumbs-up-if-you-like-the-models.-thanks!-8frvi03of">Please give a thumbs up if you like the models. Thanks! 👍👍</h2><h3 id="main-trigger-word-&quot;kda-ahr1&quot;-tvarp37r5">Main Trigger Word - "KDA Ahr1"</h3><h2 id="i-have-plans-for-more-characters.-hn1qs94pi">I have plans for more characters.</h2>The additional ",,,,," commas are not required. I separated the tags with several commas. "KDA Ahr1" is the primary tag; however, you can select the other tags based on your needs. Use the tags you need; you don't have to use them all.<h3 id="i-had-decent-results-with-a-lora-strength-of-0.55.-however-these-might-not-be-effective-for-everyone-so-i-recommend-a-minimum-of-0.5-a-maximum-of-1.0-and-a-balanced-0.6.-p42jxe846">I had decent results with a LoRA Strength of 0.55. However, these might not be effective for everyone, so I recommend a minimum of 0.5, a Maximum of 1.0, and a Balanced 0.6.</h3><h3 id="i-used-a-refiner-with-the-latent-upscale-with-a-model-to-get-better-image-details-in-case-others-try-to-use-the-same-prompt.-9li7qtpyp">I used a Refiner with the latent upscale with a model to get better image details in case others try to use the same prompt.</h3>KDA Ahri's experimental LoRA.I'm not good with LoRA training, but KDA Ahri is my favorite fictional character, and I really needed that Pony version of the face in Illustrious.In case you see additional versions in the future, I'm still testing on roughly 30 epochs to determine which produces better results because I made three distinct datasets and trained 10 epochs on each. I used most of my buzz on just Queen KDA Ahri. :PI tried other IL LoRAs, but most kept giving me childlike faces, like using a score of 1 or 2 in positive in Pony; maybe my workflow and prompt setup were improper for them, so I created one that works with my desired prompting.

KDA Ahri League of Legends + Baddest Ahri

This one was trained on a variety of image datasets. Depending on prompting, it can produce better or worse results than the other ones.

Files with SHA256: e0cad8ac259778b5b2794ed83ab48606ff106999933678232b977ddb6d4ddaac