Read before use!

This Lora was created with 100 male images generated on Bing Images between the years 2023 and early 2024, all images were upscaled to 4K.

I created this Lora because I could no longer create images in this style using the current Bing Image dalle-3, even using exactly the same prompt as before, the quality dropped a lot.

This Lora is tricky to use. I think it is possible to create females, but I haven't tested it. it will depend a lot on the model you are using. It tends to give a painting/pen and ink/ballpoint pen effect to the images. All the images used to create this Lora were in realistic anime or manhwa style. All images used were SFW as I used Bing Images to create them.

Use the following prompt to give you an idea of how to use it and get the most out of Lora (weight 0.7<>2):

"(pen and ink drawing), (orange white black color scheme only), male focus,solo, man, ripped detailed abs toned muscular, (in a beach with pirate ship, background),from bottom view perspective, doing Resting chin on fist pose, delighted expression , (Long black hair with hair on the eye), orange eyes), hyper detailed face, smooth skin, extremely detailed 8K, smooth, high resolution, ultra quality, highly detail eyes, highly detail mouth, perfect eyes, both eyes are the same,, devilayres style,drawn on style, sketched style, visible shading strokes, detailed shading, painting shading, high constrast (masterpiece , best quality , amazing quality , detailed background , intricate details , intricate , absurdres , extremely aesthetic , newest:1 . 2)"

"(pen and ink drawing), (red white black color scheme only), male focus,solo, man, ripped detailed abs toned muscular, (in a beach with pirate ship, background),from bottom view perspective, doing sitting pose, smirk expression , (Long black hair with hair on the eye), red eyes), hyper detailed face, smooth skin, extremely detailed 8K, smooth, high resolution, ultra quality, highly detail eyes, highly detail mouth, perfect eyes, both eyes are the same,, devilayres style,drawn on style, sketched style, visible shading strokes, detailed shading, painting shading, high constrast (masterpiece , best quality , amazing quality , detailed background , intricate details , intricate , absurdres , extremely aesthetic , newest:1 . 2)"

I strongly recommend you to use Lora "Sketch Semi-Realistic Style (weight 0.7)" and "Outline / 轮廓线 noobai (weight -2)", they are not necessary, but in testing all the images I created using them I thought the output was better. The Lora weight depends on your taste, I like 0.7 for images where I want the effect to be lighter and 2 for images with a stronger effect. Start with 0.7 and gradually increase to see the different outputs.

This lora was trained based on illustrious, please use a model based on it and even then depending on the model the result may not be good, its effect will depend a lot on the model you are using.

For 2D Anime add the following to the prompts:

Positive Prompt: (2D anime, anime screencaps:1.3) Negative Prompt: (realistic,photo,photography,RAW,3D:1.3)

I used VXP_Illustrious a lot because of the models I tested it had the best results and it's cheap.

So I recommend these: VXP_illustrious 1.88 - work VXP_illustrious 1.9 - work VXP_(Illustrious_v2.0 based) - work

Personally, I like the VXP_Illustrious 1.88 result better.

CFG 4.5 Steps 12 DPM++ SDE Karras

Add (tattoos:1.5) in the negative prompt if you want to remove tattoos and marks on the body

Test and see what works best for you! (Sorry for any mistakes, english is not my main language)