sushan98/nepal-lora-test

🏔️ Nepal LoRA Fine-Tune (Stable Diffusion v1.5)

Trained by Sushan Ghimire (@sushan98).

This LoRA is fine-tuned on Stable Diffusion v1.5 (stable-diffusion-v1-5/stable-diffusion-v1-5) to generate high-fidelity imagery representing the culture, architecture, traditional attire, landscapes, and landmarks of Nepal.

---

📦 Model Details

• Base Model: stable-diffusion-v1-5/stable-diffusion-v1-5
• Weights Format: Safetensors (pytorch_lora_weights.safetensors)
• Task: Text-to-Image Generation
• Author: Sushan Ghimire

---

🚀 Usage with diffusers

import torch
from diffusers import StableDiffusionPipeline

# Load base SD 1.5 model
pipe = StableDiffusionPipeline.from_pretrained(
    "stable-diffusion-v1-5/stable-diffusion-v1-5",
    torch_dtype=torch.float16
).to("cuda")

# Load Nepal LoRA weights from sushan98
pipe.load_lora_weights("sushan98/nepal-lora-test")

# Generate Image
prompt = "A breathtaking panoramic view of Kathmandu Valley, traditional Newari architecture, sacred temples, golden hour lighting, 8k resolution, highly detailed"
image = pipe(prompt, num_inference_steps=30, guidance_scale=7.5).images[0]

image.save("nepal_scenery.png")

---

🖼️ Suggested Prompts & Styles

• Heritage & Architecture: "Patan Durbar Square with intricate wood-carved temple windows, prayer flags fluttering in mountain breeze, cinematic lighting"
• Traditional Portraits: "A traditional Nepali portrait in Daura Suruwal and Dhaka Topi, realistic skin texture, warm studio portrait"
• Himalayan Landscapes: "Mount Everest and Machapuchare peaks glowing at sunrise, Sherpa prayer stones, photorealistic, 8k"

---

⚖️ License

CreativeML OpenRAIL-M license (consistent with Stable Diffusion v1.5).