Trigger word: <code>DHHL</code> Weight: <code>1.0 - 1.3</code>(Can try higher/lower with)Since the dataset contains almost no human characters, I suggest describing character features in detail, like: <code>black eyes, black hair, black shirt, black skirt, upper body, clothes, light smile,</code>The dataset is full of animals, so just to be safe, you might add these to negative prompts (not required): <code>furry, animal,</code><hr />This LoRA is something I use to tweak images, but I’m not entirely sure how to describe the exact effects. It was trained using frames from the animated film The Big Bad Fox and Other Tales – I really loved its drawing style and color. The final training results are somewhat satisfying to me. It can produce some of the subtle outline strokes from the original work - exactly what I wanted.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/ef1d486a-f807-4b52-831b-f81a8dfe5e70/width=525/ef1d486a-f807-4b52-831b-f81a8dfe5e70.jpeg" />This LoRA cannot accurately reproduce any characters from the animation because I intentionally didn't label characters during the tagging process (this was an experimental design choice). The reason it's called "Additive" is simply that I noticed the images it generates alone aren't very good-looking, but when combined with other style tags or LoRAs, they might produce better results. That's why I refer to it as an additive - feel free to use it however you like =w=<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6f31750f-14fd-4a04-b440-8c979c24f106/width=525/6f31750f-14fd-4a04-b440-8c979c24f106.jpeg" />When weight exceeds 1.5, the background will be damaged. However, adding 'simple background' to suppress it enables safe use of higher weights.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/77448d4b-9d15-4658-93ee-e413403d6eb2/width=525/77448d4b-9d15-4658-93ee-e413403d6eb2.jpeg" />

Files with SHA256: 28c5584d4a771415826eb16f30b0ba1f434d6bc2664d91c7ff32e3232d89a9e1