Update!:

Anima version is released. So, I was thinking about my options for the resolution.

I could do 1024x as usual or try 1536x. But I said: "You know what? I'm gonna do both." LOL.

It's trained on both 1024x and 1536x at the same time.

This means; you can go up to higher resolutions and have better quality. Because it's trained on high-res unlike older versions.

I recommend 1216x1856 or 1856x1216.

It's also trained on 1024x, you can generate at usual resolutions as well.

OLD (Illustrious):

Added Illustrious version.

OLD (Pony):

This lora is a mixture of works of the two Naruto Bomb artists, Raykus and Cyberunique.

The actual title is [Raykus and Cyberunique Style] but I thought this was too long, so I shortened it. Why not Naruto Bomb? Because it's like 90% Naruto and 10% something else.

If you want to make a specific character, it's highly recommended to use an additional lora for that specific character.

This lora was made with intention to create a style lora, not characters. Although it definitely recognizes some characters. Especially Uzumaki Naruto (To the point you don't need an extra lora for him)

I tried to cut watermark as much as possible without losing diversity, overall it depends on RNG whether or not to add watermark. When using ADetailer to improve faces, set "Detection model confidence threshold" to at least 0.5 in the Detection section. This is necessary to not make the extension confused.