SDXL Lightning (Fast Model)

SDXL Lightning is our optimized image generation model that provides rapid results while maintaining good quality. Based on ByteDance's optimizations of SDXL, this model is perfect for rapid prototyping and iterative design work.

Model Overview

Technical Specifications

• Provider: ByteDance
• Base Architecture: Stable Diffusion XL
• Resolution: Up to 1024x1024
• Default Steps: 10 (optimized)
• Default Guidance: 7.5 (optimized)
• Key Feature: Fast generation (2x faster than standard)

Basic Usage

import { Neuredge } from '@neuredge/sdk';

const neuredge = new Neuredge({
  apiKey: 'your-api-key'
});

const image = await neuredge.generateImage({
  model: '@cf/bytedance/stable-diffusion-xl-lightning',
  prompt: 'A quick concept sketch of a modern office space',
  width: 1024,
  height: 1024,
  num_steps: 10,
  guidance: 7.5
});

Real-World Applications & Examples

1. Rapid Design Iteration

async function generateDesignConcepts(concept, variations = 4) {
  const concepts = await Promise.all(
    Array(variations).fill(0).map((_, i) => 
      neuredge.generateImage({
        model: '@cf/bytedance/stable-diffusion-xl-lightning',
        prompt: \`${concept.base_prompt}, 
        variation ${i + 1}, ${concept.style},
        quick concept visualization\`,
        width: 1024,
        height: 1024,
        num_steps: 10
      })
    )
  );
  
  return concepts;
}

Use Cases:

• Design brainstorming
• Concept exploration
• Mood board creation
• Quick iterations
• Client presentations

2. UI/UX Prototyping

async function generateUIElement(component) {
  const response = await neuredge.generateImage({
    model: '@cf/bytedance/stable-diffusion-xl-lightning',
    prompt: \`UI design for ${component.type},
    ${component.style} style, clean and minimal,
    showing ${component.state} state,
    interface design, digital product design\`,
    width: 1024,
    height: 1024,
    num_steps: 10
  });
  
  return response;
}

Use Cases:

• Interface mockups
• Design system elements
• Component variations
• Interactive prototypes
• UI pattern libraries

3. Content Draft Generation

async function generateContentDraft(campaign) {
  const draft = await neuredge.generateImage({
    model: '@cf/bytedance/stable-diffusion-xl-lightning',
    prompt: \`Quick draft for ${campaign.type},
    featuring ${campaign.elements.join(', ')},
    ${campaign.style} style, rough concept,
    for review purposes\`,
    width: 1024,
    height: 1024,
    num_steps: 8  // Even faster for initial drafts
  });
  
  return draft;
}

Use Cases:

• Social media drafts
• Content planning
• Campaign concepts
• Draft visualizations
• Quick mockups

Integration Examples

Express.js Quick Mockup Service

import express from 'express';
import { Neuredge } from '@neuredge/sdk';

const app = express();
app.use(express.json());

const neuredge = new Neuredge({
  apiKey: process.env.NEUREDGE_API_KEY
});

app.post('/quick-mockup', async (req, res) => {
  try {
    const { concept, variations = 3 } = req.body;
    
    // Generate multiple variations quickly
    const mockups = await Promise.all(
      Array(variations).fill(0).map(() => 
        neuredge.generateImage({
          model: '@cf/bytedance/stable-diffusion-xl-lightning',
          prompt: concept,
          width: 1024,
          height: 1024,
          num_steps: 10
        })
      )
    );

    // Convert all to base64
    const base64Images = await Promise.all(
      mockups.map(async img => {
        const buffer = await img.arrayBuffer();
        return Buffer.from(buffer).toString('base64');
      })
    );

    res.json({
      mockups: base64Images.map(base64 => 
        \`data:image/png;base64,\${base64}\`
      )
    });
  } catch (error) {
    res.status(500).json({ error: error.message });
  }
});

Next.js Live Design Tool

// pages/api/live-preview.js
import { Neuredge } from '@neuredge/sdk';

const neuredge = new Neuredge({
  apiKey: process.env.NEUREDGE_API_KEY
});

export default async function handler(req, res) {
  if (req.method !== 'POST') {
    return res.status(405).json({ message: 'Method not allowed' });
  }

  try {
    const { design, style } = req.body;

    // Quick preview generation
    const preview = await neuredge.generateImage({
      model: '@cf/bytedance/stable-diffusion-xl-lightning',
      prompt: \`${design} in ${style} style, quick preview\`,
      width: 1024,
      height: 1024,
      num_steps: 10
    });

    const buffer = await preview.arrayBuffer();
    res.setHeader('Content-Type', 'image/png');
    res.send(Buffer.from(buffer));
  } catch (error) {
    res.status(500).json({ error: error.message });
  }
}

Best Practices

1. Prompt Engineering

   • Keep prompts concise
   • Focus on key elements
   • Use clear descriptions
   • Specify design style

2. Iteration Strategy

   • Generate multiple variants
   • Start with minimal steps
   • Refine promising concepts
   • Use parallel generation

3. Performance

   • Implement caching
   • Use batch requests
   • Monitor response times
   • Handle timeouts

4. Workflow Integration

   • Set up feedback loops
   • Automate common tasks
   • Track iterations
   • Maintain version history

Parameters Guide

Parameter	Range	Description
num_steps	8-12	Optimized for speed
guidance	6.0-8.0	Balance of creativity/accuracy
width	512-1024	Image width in pixels
height	512-1024	Image height in pixels

When to Use

✅ Ideal For:

• Rapid prototyping
• Design exploration
• Concept iteration
• Quick previews
• Draft generation

❌ Consider Alternatives When:

• Final deliverables needed
• Maximum quality required
• Client-ready assets
• Production materials

Getting Started

1. Install the SDK:

npm install @neuredge/sdk
# or
pip install neuredge-sdk

2. Set up your environment:

NEUREDGE_API_KEY=your-api-key

3. Initialize the client:

import { Neuredge } from '@neuredge/sdk';

const neuredge = new Neuredge({
  apiKey: process.env.NEUREDGE_API_KEY
});

Learn more:

• Quick Start Guide
• JavaScript SDK
• Python SDK