ArtificialGuyBR

Home / Blog / Training LoRAs

ColoringBook.Redmond 2.1V – Coloring Book LoRA for SD 2.1

LoRA for Stable Diffusion 2.1 that generates clean line-art coloring book illustrations. Trigger with Coloring Book and ColoringBookAF.

1 sources cited Training LoRAs

On this page

  1. What this model does
  2. How it was trained
  3. How to use it
  4. License
  5. Common errors and fixes
  6. FAQ
  7. Sources

ColoringBook.Redmond 2.1V – Coloring Book LoRA for Stable Diffusion 2.1

TL;DR

What this model does

ColoringBook.Redmond 2.1V is a Low-Rank Adaptation (LoRA) trained on stabilityai/stable-diffusion-2-1-base that steers generations toward a coloring-book aesthetic: bold, clean lines, minimal shading, and open areas ready for color fill. It was created by ArtificialGuyBR and released on Hugging Face.

The model card shows example outputs across a range of subjects — animals (owl, lion, fat cat), characters (superhero, pirate boy, portrait of an asian woman), patterns (mandala), and vehicles (Ferrari) — all sharing the same consistent line-art style.

How it was trained

The author does not publish the exact training dataset size, learning rate, or step count on the model card.

How to use it

With 🧨 diffusers (Python)

from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained(
    'stabilityai/stable-diffusion-2-1-base',
    torch_dtype=torch.float16
).to('cuda')

pipeline.load_lora_weights(
    'artificialguybr/coloringbook-redmond-2-1v-coloring-book-lora-for-freedomredmond-sd-2-1',
    weight_name='ColoringBookRedmond21V-FreedomRedmond-ColoringBook-ColoringBookAF.safetensors'
)

image = pipeline('A cute owl, Coloring Book, ColoringBookAF').images[0]
image.save('owl-coloring.png')

Prompt format

Always include both trigger tokens:

<your subject>, Coloring Book, ColoringBookAF

Examples from the model card:

License

The model uses a bespoke LoRA license that permits:

It does not allow relicensing under a different license. Full terms at the license link.

Common errors and fixes

Error | Cause | Fix | Source

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

"LoRA weights not found" | Wrong weight_name | Use exact filename: ColoringBookRedmond21V-FreedomRedmond-ColoringBook-ColoringBookAF.safetensors | Model card Style not applying | Missing trigger words | Always include both Coloring Book and ColoringBookAF | Model card CUDA OOM | fp16 + large batch | Reduce batch size or use CPU offload | diffusers docs

FAQ

Q: Does this work with SDXL or SD 1.5? A: No. It was trained specifically on SD 2.1 base. Using it on other base models will produce poor results.

Q: Can I merge this LoRA into the base model? A: Yes, diffusers supports fuse_lora() for permanent merging. See the diffusers LoRA docs.

Q: Is there a Civitai page? A: Not linked from the Hugging Face model card. The model appears to be HF-only.

Q: What resolution works best? A: SD 2.1 base is natively 512×512. For higher resolutions, use hi-res fix or upscale after generation.

Sources