<h1 id="obligatory-disclaimer:-this-lora-is-intended-to-be-used-alongside-style-loras-andor-artist-style-tags-to-bring-out-its-full-potential.-all-example-images-use-wai-anima-and-the-&quot;haku89&quot;-artist-style-tag.">Obligatory Disclaimer: This LoRA is intended to be used alongside style LoRAs and/or artist style tags to bring out its full potential. All example images use WAI-Anima and the "haku89" artist style tag.</h1><h2 id="if-you-don't-see-all-4-example-images-try-changing-browsing-level-to-&quot;r&quot;-or-lower.">If you don't see all 4 example images, try changing Browsing Level to "R" or lower.</h2><h3 id="comments-and-gallery-posts-are-encouraged-and-appreciated-regardless-of-their-content!-(within-tos-of-course.)">Comments and gallery posts are encouraged and appreciated regardless of their content! (Within ToS, of course.)</h3>Start with a model strength of 1.0. If you find the model to be too stiff, try lowering it to 0.80.A character LoRA for the indie VTuber Detective Riko who uses the Dog Girl Detective premade model by SnowBear Project. Trained on a dataset of 40 images that consists largely of cut up screenshots and images generated/enhanced by ChatGPT.I've changed my LoRA training settings once more in hopes of better flexibility. I've swapped back to Anima Base since I think the generations look marginally better and background generation appears to be significantly nicer. I've kept the halved step count Prodigy optimizer though. That being said, it's still not perfect and the are definitely still issues. Most notably, the model sometimes confuses the magnifying glass present in the training data as a monocle despite it being clearly labeled as a magnifying glass and no monocle mentioned in any of the training prompts. It's ultimately not a big deal and doesn't happen often. The model also sometimes doesn't generate the fur pattern on her tail correctly sometimes but this is a very minor issue. The below and attached triggers are general suggestions. You can almost certainly get better and more consistent results by more specific prompting. I'm just lazy. Play around with it to see what works for you. Remember, Anima loves natural language so converting the listed triggers into natural language can result in more consistent generation.Detailers are highly recommended for the face and eyes.Manual inpainting or image refinement through ChatGPT are recommended if you're looking to generate small details.Probably won't be updated.All example images were generated without a refiner step. You can probably make your generations look much better with improved workflows.Triggers:DetectiveRiko, short hair, single braid, blonde hair, white hair strand, pink inner hair, oval eyebrows, yellow eyes, large breasts, dog ears, light brown tail with white tip,bone hair clip, brown plaid detective hat, brown detective jacket, white shirt, brown neck bow with brown jewel, dark drown tie, black belt, brown skirt, brown waist bag, half pleated skirt, half plaid pencil skirt, black thigh strap, white thighhighs, brown shoes,

Files with SHA256: 3a730e8cbc16c3bf8c979f73f4bb2fdc710a11bd5e22cf3795f73a5de5817c58