ArtificialGuyBR

Home / Blog / Training LoRAs

SD 1.5 Coloring Book LoRA: Printable Line Art in One Trigger

ColoringBookAF turns any SD 1.5 prompt into minimalist coloring book line art. Trigger, settings, and diffusers setup.

1 sources cited Training LoRAs

On this page

  1. What the model does
  2. Training details
  3. How to use it
  4. Where it fits
  5. Common errors and fixes
  6. FAQ
  7. Sources

SD 1.5 Coloring Book LoRA: Printable Line Art in One Trigger

Coloring Book Redmond 1.5V is a Stable Diffusion 1.5 LoRA that turns any subject into minimalist coloring book line art, activated with a single trigger word. No ControlNet, no post-processing — just clean black line drawings ready for printing or digital coloring.

Portrait rendered as minimalist coloring book line art

Garden scene as uncolored coloring book line art

What the model does

The model card's built-in examples show the same visual recipe across very different subjects: a portrait, a lion, a garden, a tall building, and a kids' house. All come back as minimalist line art — thin, black, uncolored outlines on a plain background, the same look as a printed coloring page. The model card demonstrates consistency across humans, animals, architecture, and scenery with the phrase "minimalist, Coloring Book" plus the trigger word.

Training details

The file shipped on Hugging Face is a Safetensors LoRA: ColoringBookRedmond15V-LiberteRedmond-ColoringBookAF.safetensors. The name still carries the "LiberteRedmond" label but the weights load on the standard SD 1.5 base, per the model card.

How to use it

ComfyUI or A1111: load the .safetensors file as a LoRA (weight ~0.8–1.0), then prompt with your subject, minimalist, Coloring Book, ColoringBookAF. Drop the LoRA weight if line detail gets too heavy.

Python / diffusers:

from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained(
    "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16
).to("cuda")
pipeline.load_lora_weights(
    "artificialguybr/coloringbook-redmond-1-5v-coloring-book-lora-for-liberteredmond-sd-1-5",
    weight_name="ColoringBookRedmond15V-LiberteRedmond-ColoringBookAF.safetensors",
)
image = pipeline("A kids house, minimalist, Coloring Book, ColoringBookAF").images[0]

More on weighting and fusing LoRAs is in the diffusers LoRA loading docs.

Where it fits

Coloring book pages, printable activity sheets, kids' book illustrations, and any workflow that needs consistent uncolored line art across many subjects. The one-trigger design makes it easy to swap subjects while keeping a uniform page style. For other style work by the same author, browse more Stable Diffusion LoRA models or the coloring and drawing style collection.

Common errors and fixes

Error | Cause | Fix | Source

|---|---|---|---

Output is grayscale-filled, not line art | LoRA not triggered | Add ColoringBookAF to the prompt | Model card trigger words Lines too heavy / mush | LoRA weight too high | Lower the weight toward 0.7–0.8 | LoRA practice Style doesn't apply | Wrong base model | Use the SD 1.5 base, not SDXL or 2.x | Model card base_model Artifacts around text | Tokenizer out of distribution | Keep the trigger word plain, no emphasis syntax | Model card usage

FAQ

Is this an SDXL model? No — it is trained for Stable Diffusion 1.5 (runwayml/stable-diffusion-v1-5). The "1.5V" in the name refers to the 1.5-version family, and the model card's diffusers snippet uses the SD 1.5 base.

Do I need to credit the author? No — the license link allows use without credit, commercial use with rent-style terms, and derivatives.

Is a ControlNet needed? No. The LoRA outputs plain coloring book pages directly; no edge-detection pass or sketch conditioning is documented.

Sources