Studio Ghibli Redmond LoRA: Ghibli Style on SD 2.1 / SDXL
Generate Studio Ghibli-style illustrations with this LoRA for Stable Diffusion. Trigger words, settings, and example images.
On this page
Studio Ghibli Redmond LoRA: Ghibli Style on SD 2.1 / SDXL
TL;DR
- LoRA that adds Studio Ghibli aesthetics to Stable Diffusion 2.1 or SDXL
- Trigger words:
Studio Ghibli,StdGBRedmAF - Non-commercial fan project — character rights belong to Studio Ghibli
What This Model Does
Studio Ghibli Redmond is a LoRA (Low-Rank Adaptation) that teaches Stable Diffusion to generate images in the unmistakable Studio Ghibli style: soft watercolor backgrounds, expressive character designs, magical fantasy environments, and the characteristic color palettes seen in films like Spirited Away, Princess Mononoke, and My Neighbor Totoro.
The model was created by ArtificialGuyBR with GPU compute sponsored by Redmond.AI, trained on a large dataset of Ghibli-style artwork.
Base Model Discrepancy
There's a conflict between the two model hosts: Source | Base Model Claimed
|--------|-------------------
HuggingFace | SDXL 1.0 (stabilityai/stable-diffusion-xl-base-1.0) Civitai | SD 2.1 (Freedom Redmond SD 2.1) The model filename includes "2.1V" which strongly suggests SD 2.1 lineage. The HuggingFace metadata may reflect where the model was uploaded rather than where it was trained. Recommendation: try loading it on both SD 2.1 and SDXL bases and compare results.
Trigger Words & Usage
Instance prompt / trigger token: StdGBRedmAF
Style trigger: Studio Ghibli
Example prompt:
StdGBRedmAF, Studio Ghibli, magical forest scene, totoro-like creature, soft lighting, watercolor style, detailed background, masterpieceRecommended settings:
- LoRA weight: 0.7–1.0 (start at 0.8)
- Steps: 20–30
- CFG: 7–9
- Sampler: DPM++ 2M Karras or Euler a
- Resolution: 512×512 (SD 2.1) or 1024×1024 (SDXL)
License & Disclaimer
- License: CreativeML OpenRAIL-M
- Usage: Non-commercial, fan-made creation for entertainment only
- Rights: All characters and IP belong to Studio Ghibli
- Contact: [email protected] for concerns
This is not an official Studio Ghibli product. Commercial use of generated images featuring Ghibli characters may infringe copyright.
Common Errors and Fixes
Error | Cause | Fix
|-------|-------|-----
LoRA not applying | Wrong base model | Test both SD 2.1 and SDXL bases Weak style transfer | LoRA weight too low | Increase to 0.8–1.0 Burned/oversaturated images | CFG too high | Lower CFG to 7–8 Trigger word ignored | Prompt formatting | Place StdGBRedmAF at start of prompt
FAQ
Q: Which base model should I use? A: The name suggests SD 2.1. Try SD 2.1 first (Freedom Redmond or stock SD 2.1), then SDXL if results are poor.
Q: Can I use this commercially? A: No. The license and disclaimer explicitly state non-commercial fan use only. Ghibli character rights are not cleared.
Q: Does it work with ComfyUI / A1111 / diffusers? A: Yes, it's a standard .safetensors LoRA compatible with all major UIs.
Q: What's the file size? A: 104.63 MB (SafeTensor format).
Sources
- HuggingFace model card: https://huggingface.co/artificialguybr/StudioGhibliRedmond
- Civitai model page: https://civitai.com/models/210452
- Redmond.AI compute sponsor: http://Redmond.AI
- Demo space: https://huggingface.co/spaces/artificialguybr/artificialguybr-demo-lora
Internal links: [SD 2.1 LoRA guide, Anime style LoRAs, Non-commercial AI art]