ArtificialGuyBR

Home / Blog / Training LoRAs

StickersRedmond SD 2.1 LoRA - Sticker Style Generation

LoRA for Freedom Redmond SD 2.1 generating glossy stickers on white backgrounds. Trigger words, recommended settings, and usage guide.

By ArtificialGuyBR Published Updated 2 sources cited Training LoRAs

TL;DR

  • Trigger with Sticker or Stickers in your prompt.
  • Load the LoRA into Freedom Redmond SD 2.1 at ~0.6-1.0 strength.
  • Outputs glossy die-cut sticker style with clean white backgrounds.
  • License: CreativeML OpenRAIL-M.

What This Model Does

StickersRedmond is a LoRA (Low-Rank Adaptation) fine-tuned on Freedom Redmond SD 2.1, a popular fine-tune of Stable Diffusion 2.1. The author, ArtificialGuyBR, trained it on a curated dataset of sticker artwork with GPU compute sponsored by Redmond.AI.

The result is a specialized adapter that biases SD 2.1 toward a glossy, die-cut sticker aesthetic: bold saturated colors, sharp clean outlines, and a consistent white background — exactly what you'd expect from physical sticker sheets.

Example sticker output showing glossy style with white background

Trigger Words

Use these in your prompt to activate the LoRA style:

  • Sticker (singular)
  • Stickers (plural)

The instance prompt used during training was Sticker, Stickers. Include one or both for best results.

How to Use

Automatic1111 WebUI

  1. Place StickersRedmond.safetensors in models/Lora/
  2. Add <lora:StickersRedmond:0.8> to your prompt (adjust weight 0.6-1.0)
  3. Include trigger word: Sticker, cute cat, white background

ComfyUI

diffusers (Python)

from diffusers import StableDiffusionPipeline
import torch

pipe = StableDiffusionPipeline.from_pretrained(
    "runwayml/stable-diffusion-v2-1",
    torch_dtype=torch.float16
).to("cuda")
pipe.load_lora_weights("artificialguybr/StickersRedmond")

image = pipe("Sticker, cute robot, white background", num_inference_steps=30).images[0]

Parameter | Value

|---|---

Base Model | Freedom Redmond SD 2.1 (or SD 2.1 base) LoRA Weight | 0.6 - 1.0 Resolution | 768x768 (SD 2.1 native) Sampler | DPM++ 2M Karras / Euler a Steps | 20 - 30 CFG Scale | 7 - 9 Another sticker example showing character consistency

Common Errors and Fixes

Error | Cause | Fix

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

Weak sticker style | LoRA weight too low | Increase to 0.8-1.0 Dirty background | Missing trigger word | Add Sticker or Stickers to prompt Oversaturated colors | CFG too high | Lower CFG to 7-8 Model not loading | Wrong base model | Use Freedom Redmond SD 2.1 or SD 2.1 base

FAQ

Q: Does this work with SDXL? A: No, this LoRA is trained specifically for SD 2.1 (Freedom Redmond). For SDXL, use the StickersRedmond SDXL variant.

Q: What license applies? A: CreativeML OpenRAIL-M, inherited from the Stable Diffusion 2.1 base model.

Q: Can I use this commercially? A: Yes, under the CreativeML OpenRAIL-M license terms — see the license for details.

Q: Why are my stickers not on white backgrounds? A: Ensure you're using the trigger words Sticker or Stickers. The model was trained with white backgrounds as part of the sticker aesthetic.

Sources

  • Hugging Face model card: https://huggingface.co/artificialguybr/StickersRedmond
  • Civitai model page: https://civitai.com/models/210876
  • Author portfolio: https://artificialguy.com/
  • Redmond.AI (GPU sponsor): https://redmond.ai/