Core Concepts

Using Diffusers Library

from diffusers import StableDiffusionPipeline
import torch

# Load model
pipe = StableDiffusionPipeline.from_pretrained(
    "stabilityai/stable-diffusion-2-1",
    torch_dtype=torch.float16
).to("cuda")

# Generate image
image = pipe(
    prompt="a cat wearing sunglasses, digital art",
    negative_prompt="blurry, low quality",
    num_inference_steps=50,
    guidance_scale=7.5,
    height=512,
    width=512
).images[0]

image.save("output.png")

Prompt Tips

• Be specific and descriptive
• Add style keywords (oil painting, 3D render, etc.)
• Use negative prompts to avoid unwanted elements
• Include quality modifiers (highly detailed, 4k, etc.)
• Reference artists or art styles
• Adjust CFG scale if image doesn't match prompt

Best Practices

• Start with 512x512 for speed
• Use 20-50 steps (more steps = diminishing returns)
• Save seed for reproducibility
• Use img2img for variations