Sayaka Maizono, the pasta princess herself and the ultimate pout and pop idol from Danganronpa 1 (Trigger Happy Havoc) and Danganronpa the Animation. This lora should let you create images with Sayaka in a few different art styles and clothing options.

IMPORTANT: This sdxl lora is trained on the same data as my original flux Sayaka Maizono lora. The process has been surprisingly smooth up until now. I found a number of issues with this lora. Thankfully I know where they stem from and intend to make an updated lora at some point in the near future. It'll give me a chance to update the flux lora as well so that they're more in line with my later danganronpa loras. However, for the moment you'll want to use some band-aid solutions. I'll have a list of the issues near the bottom of this writeup in the "bugs" section.

One of the main reasons why I wanted to "port" this over is that I trained using natural language for the captions rather than tags. So natural language prompts will probably give you the best results with this lora. However, it does fairly well with a combination of natural language and tags. I tossed some tags at it with random wildcard values and was surprised by how well some of them came out. You should find some examples in the gallery.

I trained on a few specific image types which you can replicate, to varying levels of success. I tried to include a range of the prompt options in the images so you can easily compare and contrast. But these are the major options.

Danganronpa the Animation: Use 'This is an anime screencap from Danganronpa The Animation' in the prompt to emulate the art style from the anime series Danganronpa the animation.

Danganronpa 3 (anime): Use "This is an anime screencap from Danganronpa 3."

Sprites from Danganronpa: Mention 'danganronpa sprite style' in the prompt to try emulating the art style of Danganronpa. For example, 'this image is in the danganronpa sprite style'. Despite the name, I trained on full screenshots, with and without the GUI, with this description along with character sprite rips. So in theory it should be able to 'danganronpafy' a full image.

Clothing: You can try different clothing options with prompts like the following:

Standard outfit: The school uniform Sayaka wears for most of Danganronpa 1. "(((blue socks))), Sayaka Maizono wearing her uniform from Danganronpa 1. She has her long blue hair down. She is wearing the white and blue Hope's Peak Academy uniform with a bright white top. Her sleeves reach her wrists. She's wearing a light pink ribbon. She has a short, pleated, blue skirt. Sayaka Maizono wears navy blue thigh-high socks. Her footwear is dark brown Mary Janes shoes."

Swimsuit: The bikini Sayaka wears in Danganronpa S. "Sayaka Maizono wearing her swimsuit from Danganronpa S. Her swimsuit is a light blue bikini with a white ruffle, and a teal sarong tied around her waist."

Locations: I didn't try to train on any specific locations, but some of them appeared enough in the training data that you might be able to improve your results by specifying areas from Danganronpa 1.

Training material: The majority of the training data came from Danganronpa 1, Danganronpa S, and Danganronpa the Animation. Additionally there's a smattering of Danganronpa 3 and as much official supplementary material added in as possible. The game-related data includes all sprites and a handful of in-game screenshots. LLM captioned, then hand edited.

Bugs: I know where the bugs stem from and do intend to create an updated lora eventually. However, until then, there are some usability issues with this lora. In particular the color of Sayaka's uniform tends to need extra guidance. And her socks in particular really tend to veer away from what they should be. Thankfully, just being careful with the prompt is enough to keep things more or less on track. With the socks prefacing the prompt with (((blue socks))), and with her clothing just putting emphasis on 'white' in the description. Again, an update is planned to fix that necesity.