<h3 id="short-version:-version-2-is-preferable-over-v1-and-tends-to-be-a-net-improvement.-v2-alternative-is-based-on-some-images-the-model's-commissioner-gave-me.-i'd-say-pick-either-v2-or-v2-(alternative).-0anh48v7o">Short version: Version 2 is preferable over V1 and tends to be a net improvement. V2 alternative is based on some images the model's commissioner gave me. I'd say pick either V2 or V2 (alternative).</h3>This is the Illustrious version of <a target="_blank" rel="ugc" href="https://civarchive.com/models/962018/commission-miaka-yuuki-fushigi-yuugi">my Miaka Pony LoRA</a>. It uses the same dataset, so similar images should be possible. Obligatory self-promotion: consider supporting or commissioning me at <a target="_blank" rel="ugc" href="https://www.patreon.com/MetalChromeX">https://www.patreon.com/MetalChromeX</a> or <a target="_blank" rel="ugc" href="http://www.fiverr.com/s/VY34o4m">http://www.fiverr.com/s/VY34o4m</a> so I can afford an Nvidia 5090 when I can buy one priced at or near the MSRP (screw you, scalpers!)!Update (4-9-25): Based on feedback from the model's commissioner, I made some tweaks to the V2 model; I'm calling this version V2 (alternative). V2 was trained only on anime images, while this alternative version was trained on some images the commissioner gave me.Update (4-6-25): I added an improved V2 with more and better images in the dataset and removed some questionable quality images. I'd recommend V2 over V1.

Files with SHA256: 1c7520b28c824039e7fb01ac53fec246a5acc200baadc463698ae4a0c855f0e5