GFPGAN

gfpgan

GFPGAN is a photo Enhancer AI model for improving overall photo quality and resolution.

L40S 45GB

Fast Inference

REST API

Try in Console API Docs Examples

Model Information

Response Time~5 sec

StatusActive

Version

0.0.1

Updated10 days ago

API Reference

View Full Documentation

Prerequisites

Create an API Key from the Eachlabs Console
Install the required dependencies for your chosen language (e.g., requests for Python)

API Integration Steps

1. Create a Prediction

Send a POST request to create a new prediction. This will return a prediction ID that you'll use to check the result. The request should include your model inputs and API key.

import requests
import time

API_KEY = "YOUR_API_KEY"  # Replace with your API key
HEADERS = {
    "X-API-Key": API_KEY,
    "Content-Type": "application/json"
}

def create_prediction():
    response = requests.post(
        "https://api.eachlabs.ai/v1/prediction/",
        headers=HEADERS,
        json={
            "model": "gfpgan",
            "version": "0.0.1",
            "input": {
  "img": "your_file.image/jpeg",
  "scale": "2",
  "version": "v1.4"
}
        }
    )
    prediction = response.json()
    
    if prediction["status"] != "success":
        raise Exception(f"Prediction failed: {prediction}")
    
    return prediction["predictionID"]

2. Get Prediction Result

Poll the prediction endpoint with the prediction ID until the result is ready. The API uses long-polling, so you'll need to repeatedly check until you receive a success status.

def get_prediction(prediction_id):
    while True:
        result = requests.get(
            f"https://api.eachlabs.ai/v1/prediction/{prediction_id}",
            headers=HEADERS
        ).json()
        
        if result["status"] == "success":
            return result
        elif result["status"] == "error":
            raise Exception(f"Prediction failed: {result}")
        
        time.sleep(1)  # Wait before polling again

3. Complete Example

Here's a complete example that puts it all together, including error handling and result processing. This shows how to create a prediction and wait for the result in a production environment.

try:
    # Create prediction
    prediction_id = create_prediction()
    print(f"Prediction created: {prediction_id}")
    
    # Get result
    result = get_prediction(prediction_id)
    print(f"Output URL: {result['output']}")
    print(f"Processing time: {result['metrics']['predict_time']}s")
except Exception as e:
    print(f"Error: {e}")

Additional Information

The API uses a two-step process: create prediction and poll for results
Response time: ~5 seconds
Rate limit: 60 requests/minute
Concurrent requests: 10 maximum
Use long-polling to check prediction status until completion

Overview

GFPGAN (Generative Facial Prior GAN) is a state-of-the-art AI model developed for high-quality face restoration in real-world scenarios. By leveraging generative adversarial networks (GANs) and facial priors, GFPGAN excels at restoring low-resolution, blurred, or damaged facial images while preserving high fidelity and naturalness. Its versatility makes it a popular choice for photo enhancement, historical image restoration, and creative applications.

Technical Specifications

Input Requirements

For best results, use images with clear facial regions and minimal obstructions.

Model Architecture:

Built on GAN architecture with pre-trained facial prior integration.
Refined loss functions to balance restoration and fidelity.

Input Requirements:

Formats: JPEG, PNG.
Resolution: Recommended input is up to 512x512 for optimal performance.

Output Features:

Restored images maintain original context and backgrounds.
Faces are enhanced with reconstructed features.

Key Considerations

Over-Restoration:

In some cases, the restored face might deviate slightly from the original.

Context Preservation:

Non-facial regions are minimally processed. Ensure the background meets the desired quality before input.

Tips & Tricks

Fine-Tune Settings:

Adjust restoration strength to balance detail enhancement and natural appearance.

Pre-Processing:

Crop images to focus on faces for better results.

Scale

1.0: Minimal enhancement, retains most original features.
1.5: Balanced restoration for moderate improvements.
2.0: High-level enhancement, ideal for heavily degraded images.
2.5+: Aggressive restoration, may introduce artifacts on high-quality inputs.

Capabilities

Creative Uses:

Enhances AI-generated faces or artistic projects with added realism.

Historical Photo Repair:

Revives old photographs for personal or archival purposes.

Detail Enhancement:

Recovers textures like skin, hair, and eyes with impressive clarity.

Face Restoration:

Repairs blurred, damaged, or low-quality facial images.

What can I use for?

Media Projects:

Restore archival images or enhance visuals for creative content.

Historical Preservation:

Digitally repair vintage photos for museums or personal collections.

AI Art Improvement:

Use as a finishing tool for AI-generated images to add detail and polish.

Personal Photo Enhancement:

Improve selfies, family portraits, and treasured memories.

Things to be aware of

Restore Vintage Photos:

Test the model on old or damaged images to witness its transformative abilities

Creative Enhancements:

Apply the model to artistic or AI-generated portraits for added depth.

Limitations

Generalization:

May struggle with extreme distortions or non-human faces.

Color Consistency:

Slight color variations may require manual correction.

Background Restoration:

Focuses primarily on faces, with less emphasis on backgrounds.

Output Format: PNG

Related AI Models

Stable Diffusion Inpainting

stable-diffusion-inpainting

Image to Image

SDXL Controlnet

sdxl-controlnet

Image to Image

Controlnet Realistic Vision V2.0

controlnet-1-1-x-realistic-vision-v2-0

Image to Image

Eachlabs Face Swap

each-faceswap-v1

Image to Image