Update: I do plan to work on a v3 in order to fix the scale/shell clothing issue in v2, as well as expand the individual character body types a little more. I also noticed v2 doesn't handle multiple characters well, so I'll be trying to fix this too.Something I noticed about Illustrious is that, even though it (or at least Wai-Illustrious) can generate many popular characters with decent accuracy, there are still plenty Loras for some of these characters that don't seem to add much.One such character is Bowser, but I wanted something a little different than the basic design. Thus I decided to make this Lora.<h1 id="bowser-(anthrohuman-forms)-8rrydc7ev">Bowser (Anthro/Human Forms)</h1>I've condensed the information on this model to make room for the Koopa Troop model (v2).There are two different styles that this lora can accomplish, a more typical anthro form using the 'Anthro Bowser' tag, and a more humanoid form using the 'Human Bowser' tag.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/b97fed3f-6cb1-4d02-9351-42538cd0d1ba/width=525/b97fed3f-6cb1-4d02-9351-42538cd0d1ba.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/fb2962d5-30f9-41d8-8c26-6cd4ef7e1bb7/width=525/fb2962d5-30f9-41d8-8c26-6cd4ef7e1bb7.jpeg" />As you can see, both forms keep the tail, shell, horns, and the green face, these don't need to be prompted and will come out by default. I haven't tested enough to see if negatives will eliminate them though.The claws are also default for the hands, but, as hands tend to get messy with most models, you may want to prompt for claws anyway. The eyes will also default to a red-orange color, and the hair will either default to red or blonde (due to some Bowsette crossover), but other colors should work as well. The ears are also part of the default style, and are harder to remove even with negative prompting, but not impossible.Most skin colors should work with this model, the yellow is just the default. The only color that seemed to have issues is green, possibly due to interference from the face color.The clothing, even though it was trained on a wider variety, tends to default to orange and black for some reason, but this can be fixed with simple prompting.<h1 id="the-koopa-troop-(anthrohuman-forms)-oymnvj23w">The Koopa Troop (Anthro/Human Forms)</h1>While this is technically v2 of the Bowser model, I decided to change the name due to the differences between the two models.This model expands beyond Bowser, and includes many other characters that help bring more variety to the mix. However, the trade-off for this is having less images per character, so it may not be as easy to get certain body types on some characters.Just like the first model, there are two main tags for styles, the 'Anthro &lt;character&gt;' and 'Human &lt;character&gt;' tags. The characters included are Bowser (of course), the Koopalings; Iggy, Larry, Lemmy, Ludwig, Roy, Morton and Wendy; the green/cartoon version King Koopa, Bowser Jr. (aged up), Boom Boom and Pom Pom, a Koopa Troopa and Koopa Paratroopa, Birdo, and Yoshi.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/b0c7cc0d-120b-47ec-91c9-681039072938/width=525/b0c7cc0d-120b-47ec-91c9-681039072938.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/b55530d0-4c98-44c7-a97e-0e5f7535d7e7/width=525/b55530d0-4c98-44c7-a97e-0e5f7535d7e7.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a5b18a4d-d0ec-4fb2-b24e-5be856de8773/width=525/a5b18a4d-d0ec-4fb2-b24e-5be856de8773.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6e5f7ba6-2e14-416c-a58f-eda004b28716/width=525/6e5f7ba6-2e14-416c-a58f-eda004b28716.jpeg" />The most notable difference from the first Bowser model is that I didn't include ears on the anthro forms, though they can be added with prompting. But pointed ears must now be prompted for in the human forms, otherwise you get mostly round ears.Each character was trained with male and female versions, but without proper prompting they may lean more toward feminine forms (usually due to a female bias in the checkpoint training).Most of these characters can be generated with just the 'Anthro/Human &lt;character&gt;' tags, and simple prompts, but some require more details, which can be seen in the prompt details of the image posts. Some of the Human forms don't generate with horns, but they can be added with prompting.These characters should also default with their specific face and skin colors, as well as their hair colors and styles, but you may need to prompt for it in some cases. This also means you can prompt for other colors as well. (Yes, you can create original characters this way)The stars on Larry's head and Morton's eye will sometimes work if you use the 'star' tag, but more often than not they either won't appear or will be messy.This model was trained for two shell types, Smooth and Spiked, but sometimes these will overlap. These can also be made in whatever color you want.Final note, sometimes the clothing will have a shell-like appearance, this is due to some crossover from the Shells on the characters' backs. I wasn't able to completely remove this, but adding "cloth" to the clothing prompts helps sometimes.Body Type Notes:The specific prompts used can be seen in the image posts, but a quick run down (these should work across any loras, as long as the checkpoint used has enough training data);'Chubby' and 'Plump' give bigger bodies, with chubby being larger. 'Muscular' and 'Toned' give similar bodies, with muscular typically being bigger, and adding 'Bara' in can make them even more so. 'Curvy' gives the typical feminine hourglass figure, though using 'wide hips' will push it even further, I believe 'Thicc' or 'Thick' gives a similar output but may push toward the chubby/plump area. 'Skinny' actually works better for most illustrious checkpoints than for either Pony or SDXL, from my testing.And of course, Femboy gives a more feminine male, and Futanari gives a female with a little extra. ;)These are just samples of possible gender and body types, but this model should work well with anything that the base models can handle.This model was trained using the Illustrious-XL v1.0 checkpoint, but all generations are made using WAI-NSFW-illustrious-SDXL v12 and Fooocus.All of my models may be used for your own training and/or merges.

Files with SHA256: 667174757c0b652bb23c0304e0b17af4d75ae1b87ed67bcdddc08d9c4333eff0