Readme

HiDream I1 Full - Advanced AI Text-to-Image Model

Transform your ideas into reality with HiDream I1 Full, a powerful open-source image generative foundation model with 17B parameters that achieves state-of-the-art image generation quality within seconds. Built for both hobbyists and production environments, this model delivers consistent, high-quality results with industry-leading performance.

Getting Started

Getting up and running with HiDream I1 Full takes just a few minutes. First, you'll need to set up your environment:

Install the client library for your preferred language:

For JavaScript/TypeScript:

npm install --save @fal-ai/client

For Python:

pip install fal-client

Configure your authentication by adding your API key:

import { fal } from "@fal-ai/client";
fal.config({
  credentials: "YOUR_FAL_KEY_HERE"
});

Core Features

The HiDream I1 Full model excels in several key areas:

Superior Image Quality Produces exceptional results across multiple styles including photorealistic, cartoon, artistic, and more. Achieves state-of-the-art HPS v2.1 score, which aligns with human preferences.

Best-in-Class Prompt Following Achieves industry-leading scores on GenEval and DPG benchmarks, outperforming all other open-source models.

Commercial Use Allowed Generated images can be freely used for personal projects, scientific research, and commercial applications.

Streaming Support Real-time generation updates with streaming capabilities for enhanced user experience.

Basic Usage

Here's a simple example showing how to use HiDream I1 Full:

import { fal } from "@fal-ai/client";

async function generateImage() {
  try {
    const result = await fal.subscribe("fal-ai/hidream-i1-full", {
      input: {
        prompt: "a cat holding a skateboard which has 'fal' written on it in red spray paint"
      }
    });
    
    console.log(result.data.images[0].url);
    return result;
  } catch (error) {
    console.error("Generation failed:", error);
  }
}

API Parameters

Input Parameters

• prompt (required): Text description for image generation
• negative_prompt (optional): What to avoid in the generation (default: "")
• image_size (optional): Object with width and height properties (default: 1024x1024)
  • width: The width of the generated image (default: 512)
  • height: The height of the generated image (default: 512)
• num_inference_steps (optional): Number of inference steps (default: 50)
• guidance_scale (optional): How closely to follow the prompt (default: 5)
• num_images (optional): Number of images to generate (default: 1)
• enable_safety_checker (optional): Whether to enable safety checking (default: true)
• output_format (optional): Output format - "jpeg" or "png" (default: "jpeg")
• loras (optional): Array of LoRA weights to apply

LoRA Configuration

{
  "loras": [
    {
      "path": "https://example.com/lora-weights.safetensors",
      "scale": 1.0,
      "weight_name": "optional_weight_name"
    }
  ]
}

Advanced Usage Options

Streaming Generation

For real-time updates during generation:

import { fal } from "@fal-ai/client";

const stream = await fal.stream("fal-ai/hidream-i1-full", {
  input: {
    prompt: "a cat holding a skateboard which has 'fal' written on it in red spray paint"
  }
});

for await (const event of stream) {
  console.log(event);
}

const result = await stream.done();

Queue Management

For production applications, use the queue API:

// Submit request
const { request_id } = await fal.queue.submit("fal-ai/hidream-i1-full", {
  input: {
    prompt: "your detailed prompt here"
  },
  webhookUrl: "https://optional.webhook.url/for/results"
});

// Check status
const status = await fal.queue.status("fal-ai/hidream-i1-full", {
  requestId: request_id,
  logs: true
});

// Get result when complete
const result = await fal.queue.result("fal-ai/hidream-i1-full", {
  requestId: request_id
});

Best Practices

Maximize your results by following these proven approaches:

Prompt Optimization Structure your prompts clearly and specifically. The more detailed and focused your input, the better the output quality will be.

Error Handling Always implement proper error handling in your code. The API provides detailed error messages to help you quickly identify and resolve issues.

Resource Management Monitor your API usage through the dashboard and implement rate limiting in your application to prevent overconsumption.

Image Size Considerations Larger images require more processing time and cost more. Use appropriate dimensions for your use case.

Model Variants

HiDream I1 is available in multiple variants:

• fal-ai/hidream-i1-full: Full-featured model with highest quality
• fal-ai/hidream-i1-dev: Development model for testing
• fal-ai/hidream-i1-fast: Optimized for speed (16 steps)
• fal-ai/hidream-i1-full/image-to-image: Image-to-image transformation

Pricing and Usage

• Cost: $0.05 per megapixel
• Commercial Use: Generated content can be used commercially
• Open Source: Released under MIT license
• Billing: Pay only for successful generations

Technical Specifications

• Parameters: 17 billion parameters
• Architecture: Sparse Diffusion Transformer with Mixture of Experts
• Base Components: Uses FLUX.1 VAE, T5-v1.1-xxl and Llama-3.1-8B text encoders
• Processing Time: Achieves results within seconds
• Output Formats: JPEG, PNG

Technical Support

We're here to help you succeed with HiDream I1 Full:

Documentation Access comprehensive documentation at docs.fal.ai

Model Information Official model details at Hugging Face

Support Channels Get assistance through support@fal.ai or developer forums

Ready to get started? Create your account at fal.ai and begin building with HiDream I1 Full tod