🎨 Indian Art Stable Diffusion Model
Generate beautiful traditional Indian art styles using simple text prompts. Fine-tuned from Stable Diffusion v1.5 on Indian cultural art themes.
🖼️ Sample Outputs
✅ What This Model Is Good At
- Madhubani / Mithila folk paintings
- Warli tribal art
- Rangoli festival designs
- Mughal miniature painting style
- Tanjore / Thanjavur classical style
- Pattachitra cloth paintings
- Indian festival scenes (Diwali, Holi, Ganesh Chaturthi)
- Traditional dancers (Bharatanatyam, Kathak)
🚀 How To Use It
In Google Colab (Easiest)
from diffusers import StableDiffusionPipeline
import torch
pipe = StableDiffusionPipeline.from_pretrained(
"Vrizzo/indian-art-sd-model",
torch_dtype=torch.float16
).to("cuda")
image = pipe(
"Madhubani painting of a peacock, intricate patterns, vibrant colors",
negative_prompt="blurry, low quality, distorted",
num_inference_steps=30,
guidance_scale=7.5
).images[0]
image.save("output.png")
🌟 Best Prompts To Try
| Style | Prompt |
|---|---|
| Madhubani | Madhubani painting of peacock, intricate patterns, Bihar folk art |
| Warli | Warli tribal art, village harvest scene, geometric figures, white on red |
| Rangoli | Diwali rangoli design, symmetrical mandala, lotus motif, vibrant colors |
| Miniature | Mughal miniature painting, royal court scene, gold accents, fine detail |
| Tanjore | Thanjavur painting, goddess Lakshmi, gold foil emboss, gemstone colors |
⚙️ Model Details
- Base model: runwayml/stable-diffusion-v1-5
- Image size: 512×512
- Recommended steps: 25–40
- Recommended guidance scale: 7–9
⚠️ Limitations
- Works best with art style keywords included in the prompt
- Photorealistic human faces are not its strength
- Best results at 512×512 resolution