Sayaka Maizono ( Danganronpa ) [SD 1.5]版本V1 (ID: 906303)

Sayaka Maizono, the ultimate pout idol and pasta princess 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 stable diffusion 1.5 lora is trained on a conversion my original flux Sayaka Maizono training data. In short, this sd 1.5 lora suffers a bit from being designed for a very, very, different framework. I wrote a system to automatically convert the flux training data to as close to the 1.5 format as possible. But I only did a small amount of manual editing afterward. Still, the testing seems to be giving acceptable results. That said, there are issues. It's probably going to be a bit unpredictable. For models, I've had the best results for this lora with ghostmix, lustermix 2D, Mistoon_Anime 3.0 and aing diffusion. Likewise, I typically used adetailer for the faces. Additionally, Danganronpa GUI elements are probably going to appear every now and then. It doesn't seem to happen 'too' often though. And you might be able to catch them with negative prompts. That said, caveats aside, I think this experiment was a success. The majority of the images I've uploaded with this lora were made through random wildcard generations for the tags which I think is a nice testament to its flexability. The ability to stick to styles like the danganronpa 3 anime style or danganronpa the animation is pretty weak as well. Finally, the tagging is messy. The price payed for automated conversion.

Unique features: The main point I try to stress in my Danganronpa loras is authenticity to the source material. It's fairly easy to get the basics of a character's style or clothing down. Characterization is a lot harder. Models have their own concepts of what a scared, or happy, or sad, or angry expression or stance would be. But a character in something like Dananronpa has a lot of careful attention put into ensuring their designs make all of those states unique. I go through every sprite, every CG, every official image I can find to caption and include them in the dataset in hope of best representing that. Whether it works out is another question, but it's what I'm aiming for and why these might be a little odd. And the unpredictability goes double for the SD 1.5 conversions. However, I think they provide a worthwhile take on the characters.

I trained on a few specific image types which you can replicate, to varying levels of success. This tends to work quite well in flux, OK with SDXL, and generally fairly poorly in SD 1.5. 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 the tags 'Danganronpa The Animation, screencap from the anime Danganronpa The Animation, anime, dr1anime' in the prompt to emulate the art style from the anime series Danganronpa the animation.

Danganronpa 3 (anime): Use "Danganronpa 3 screencap style, danganronpa 3, anime"

Sprites from Danganronpa: Use the tags 'danganronpa sprite style, Danganronpa V3 sprite style' or 'danganronpa sprite style, Danganronpa S sprite style' in the prompt to try emulating the art style of Danganronpa. 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.

Danganronpa S GUI: You can give it a try with this. "danganronpa s, danganronpa sprite style, danganronpa S sprite style, danganronpa s gui style dialog box, Danganronpa S GUI status bar, black text, Danganronpa S GUI title bar"

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

Standard outfit: "Sayaka Maizono, long blue hair, large blue eyes, Hope's Peak Academy Uniform, white sailor fuku, blue socks, white top, blue skirt, long sleeves "

Swimsuit: "Sayaka Maizono, Danganronpa Sprite Style, Blue Hair, Blue Eyes, Long Hair, Side Part, Relaxed Posture, Bikini, Light Blue, White Lace, Turquoise Sarong"

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 and tags from danganronpa or danganronpa S. In particular with 'jabberwock island', 'Danganronpa' and 'Hope's Peak'.

Training material: The majority of the training data came from Danganronpa, Danganronpa the Animation, and Danganronpa S with 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.

描述:

训练词语: