Downloads · 30 days
439
40% of all-time downloads
strkyyy/manga-ink-screentone
manga-ink-screentone is a text-to-image model from strkyyy. Use it when you need an image from a text prompt. It is set up for diffusers. The card lists the license as openrail++.
This LoRA model captures the classic black & white manga page aesthetic with clean ink line art, varied line weights, screentone dot shading, dramatic hatching for shadows, expressive characters, and traditional Japan…
Downloads · 30 days
439
40% of all-time downloads
All-time downloads
1.1K
Public
Repo size
107 MB
Likes
0
Public
Click a slice to open those files.
.safetensors93.1 MB · 87%
From the Hugging Face model README
This LoRA model captures the classic black & white manga page aesthetic with clean ink line art, varied line weights, screentone dot shading, dramatic hatching for shadows, expressive characters, and traditional Japanese manga composition.
Trigger the style using the keyword 'm4ng41nk' in your prompt. For best results, set a strength between 0.6 and 1.0 (recommended default of 0.85) and select K-LMS as your sampler with around 25 steps.
m4ng41nk A lone samurai at dusk under a stormy sky, dramatic shadows and contrast --c 11
m4ng41nk Traditional Japanese architecture with intricate details and patterns, warm lighting --s 750
m4ng41nk Close-up portrait of a determined ninja, eyes glowing in the dark, soft shadows --l medium
m4ng41nk A kendo match in full swing, intense action with dynamic lines and screentone shading --c 8
m4ng41nk Mysterious fog enveloping an ancient temple at night, eerie mood with subtle hatching --s 600
m4ng41nk Abstract composition of geometric shapes, stylized line art and bold screentones --l soft
Activate with m4ng41nk. Recommended strength: 0.9.

diffusersfrom diffusers import AutoPipelineForText2Image
import torch
pipe = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipe.load_lora_weights("strkyyy/manga-ink-screentone")
image = pipe("m4ng41nk, your prompt", num_inference_steps=30).images[0]