I am admittedly kind of a noob at this, but if you enjoy my humble work, any donations are greatly appreciated: <a target="_blank" rel="ugc" href="https://ko-fi.com/insult_to_ninjary">https://ko-fi.com/insult_to_ninjary</a>Your generosity helps me stay motivated to keep improving and to keep trying new things!Trigger Words:Main trigger: 0k4m1Helper prompts: parody, style parodyExtra help: sumi-e<ul><li>The LoRA can potentially work with no triggers, but it may not be strong/consistent.</li><li>For a stronger effect, use the main trigger: 0k4m1. You can also weight it numerically or with parentheses - I usually do (((0k4m1))).</li><li>If the (weighted) main trigger is still not enough by itself, next try one or both of the helper prompts along with the main trigger. They can be weighted as well - I usually do them all together, like (((0k4m1, parody, style parody))).</li><li>I didn't include it as an official trigger because it's usually not necessary and may be anecdotal, but adding sumi-e to the other trigger words may (or may not) help to strengthen the style if you feel it still needs a bit more of a boost.</li><li>Please note that the main trigger here is only for the Illustrious model - it won't work for the SD1.5 models. The helper prompts may or may not work for SD1.5 (I never tried them), but they are not necessary as both versions of the SD1.5 LoRA work just fine with no triggers!</li></ul>At a Glance (Illustrious):<ul><li>Base model: Illustrated style (I use WAI-illustrious-SDXL)</li><li>LoRA weight: 1 (but slowly go lower or higher if necessary)</li><li>Suggested parameters: Euler a, Automatic, 30 steps, 832x1216/1216x832, CFG 5 (but feel free to experiment!)</li><li>Hires. fix: 1.5x, use a quality upscaler (like 4x-UltraSharpV2), denoising strength 0.5 - 0.3</li><li>Extra: use ADetailer to fix faces and/or hands if necessary</li></ul>Illustrious Version Description:This LoRA produces art in a similar style to the game Okami. It may not always be super on-the-nose, but it definitely captures the vibe most of the time. Below, I'll go over some of the most common problems with this LoRA.First, outputs can be a little blurry, especially in fine details like faces, hands, and intricate background elements. You may need to use ADetailer for faces and/or hands if they don't come out good enough on the first go. I've also discovered that adding "sumi-e" to the prompt (which, I think, is the style of art used in Okami) can sometimes help a lot!In addition, the model likes to "zoom out" a bit, as it was trained with a bunch of game screenshots. The helper prompts seem to intensify this behavior further. To counteract it (if you want to), you can prompt for the specific composition you want (portrait, upper body, cowboy shot, etc.), and this usually works well enough.Finally, the LoRA seems to do much better with outdoor settings compared to indoor settings. This makes sense, as the game screenshots used in training were mostly outdoors. Thus, it's very good at capturing the style of things like trees, grass, flowers, mountains, etc. But indoor shots will be more of a crapshoot, as it's kind of just guessing at how the style should be applied to most things.For more potentially relevant information, you can check out the original SD1.5 description immediately below. Below that, there's a troubleshooting section in case you're having any problems.SD1.5 Versions Description:Style Version: This was the original LoRA I made. I think it gets a bit closer to an authentic Okami art style than the "Substance" version. The drawback is that it seems to be a bit more stubborn to work with in terms of getting your desired composition. If you're aiming for a more accurate art style and you don't mind being patient with doing multiple gens and possibly a decent amount of inpainting/post-processing, then this should be your choice. Despite the negatives, I still prefer this version.Substance Version: With this version, I tried to update the "Style" version to make it more stable, and it kind of worked, but it sacrificed a bit of authenticity in the process. So this version should give you a bit of an easier and quicker time getting good-looking images, but I feel that they look less like Okami than the "Style" version. Still, if you're not a stickler for purity, then this version will be easier to work with and still get you some cool-looking stuff!General Info: This LoRA series will generate images in an art style reminiscent of the video game Okami.I loved playing Okami, and a big reason for that was the stunning art style. With these LoRAs, hopefully you can have some success recreating it! While the LoRAs were not trained on specific characters, Amaterasu was in a fair amount of the training images, so you may have some luck prompting her. Sometimes Ammy (or some other wolf/dog-like creature) will even pop up unprompted! But mainly, these LoRAs were designed to put your own creations into an Okami-esque art style.Occasionally, the LoRAs can be a bit stubborn, especially depending on which version you use. Here are some usage tips if you're getting incoherent images, images that are too landscapey, your character is missing or very small, etc.:1. Use a VAE. I use vae-ft-mse-840000-ema-pruned. 2. If you're using Hires. fix, change your upscaler off of "Latent." I use SwinIR_4x. [Way later update: now I use 4x-UltraSharpV2] 3. Lower the weight of the LoRA. It seems to work pretty well in the 0.7 to 1 range, so if it's giving you trouble at 1, gradually decrease it until you start getting your desired composition.The LoRAs seem to work pretty well with most base models. For the most accurate effect, use a cartoon/anime model. I recommend Koji, with Doll Like Anime also being a decent choice. But if you're feeling adventurous, you can also try them with realistic or semi-realistic models. Obviously the more realistic you get, the less it will look like Okami, but it will still usually keep some of that Okami vibe and aesthetic in the form of clothing, backgrounds, etc.Troubleshooting:If you're having persistent problems with anatomy, fine detail, and/or other jank, just try your best with all the usual fixes - different prompting (including negative prompt), try a different base model, use a different VAE, try different clip skip values, adjust the LoRA weight, use adetailer, inpaint, etc.Outro:If you make anything awesome, please consider posting them to the gallery! That goes for all of my models! I really like seeing what kind of stuff you all are able to make!If you have any questions, feel free to ask! I will try my best to help! Also, if you have any ideas for future models, please don't hesitate to share them with me!

Files with SHA256: 3a02d46322aeab41d487d2d847c520dc0b7557f604d2d41c2f74dc53bcef5ca5