ByteDance with Eachlabs
ByteDance's new video model, OmniHuman, is available on Eachlabs →
Eachlabs | AI Workflows for app builders

Instant ID Generate Avatar

instant-id

Instant ID is making realistic images of real people instantly

L40S 45GB
Fast Inference
REST API
Try in ConsoleAPI DocsExamples

Model Information

Response Time~32 sec
StatusActive
Version
0.0.1
Updated28 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": "instant-id",
            "version": "0.0.1",
            "input": {
  "prompt": "a person",
  "negative_prompt": "your negative prompt here",
  "num_inference_steps": "4",
  "image": "your_file.image/jpeg",
  "pose_image": "your pose image here",
  "width": "640",
  "height": "640",
  "scheduler": "EulerDiscreteScheduler",
  "guidance_scale": "7.5",
  "pose_strength": "0.4",
  "canny_strength": "0.3",
  "enable_depth_controlnet": false,
  "depth_strength": "0.5",
  "enable_lcm": false,
  "seed": null,
  "disable_safety_checker": false,
  "enable_pose_controlnet": "True",
  "enhance_nonface_region": "True",
  "lcm_guidance_scale": "1.5",
  "ip_adapter_scale": "0.8",
  "controlnet_conditioning_scale": "0.8",
  "lcm_num_inference_steps": "5",
  "enable_canny_controlnet": false,
  "sdxl_weights": "stable-diffusion-xl-base-1.0"
}
        }
    )
    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: ~32 seconds
  • Rate limit: 60 requests/minute
  • Concurrent requests: 10 maximum
  • Use long-polling to check prediction status until completion

Overview

Instant ID Generate Avatar model leverages advanced neural architectures for generating high-quality images by combining input prompts with pose control, depth control, and conditional data. With support for a wide range of configurations, it enables users to create personalized, high-fidelity outputs while maintaining flexibility in style and structure. Instant ID Generate Avatar is designed for intuitive usability and provides fine-grained control over the generation process through an array of configurable inputs.

Technical Specifications

Architecture: Combines diffusion-based models with multi-layer conditional nets for precise image generation with Instant ID Generate Avatar.

Pre-trained Weights: Includes advanced pre-trained weights such as stable-diffusion-xl-base-1.0 and dreamshaper-xl to ensure diverse artistic outputs.

Schedulers: Multiple scheduler options, such as DEISMultistepScheduler and EulerDiscreteScheduler, are available for precise control over inference quality and speed.

Fine-Tuning Controls: Parameters such as guidance_scale, ip_adapter_scale, and controlnet_conditioning_scale provide granular control over stylistic and compositional fidelity.

Key Considerations

Prompt Quality: Clear, descriptive prompts lead to better results. Use negative_prompt to explicitly exclude undesired features.

Pose and Depth Control: Ensure pose and depth input images align with the desired output structure for effective conditioning.

Safety Checker: Enabling or disabling the safety checker impacts output filtering. Use discretion when disabling it.

Tips & Tricks

General Tips for Instant ID Generate Avatar:

  • Prompt: Use detailed and descriptive prompts for high-quality outputs. For instance, "a futuristic cityscape at sunset" yields better results than vague prompts.
  • Negative Prompt: Refine outputs by excluding unwanted elements, such as "blurry details" or "oversaturated colors."
  • Seed: Set a specific seed for reproducible results, or leave it unset for unique outputs.

Resolution:

  • width and height: Opt for resolutions that match your intended use. For example:
    • Low-resolution drafts: 640x640.
    • Final render: 2048x2048 or higher (up to 4096x4096).

Style Selection:

  • sdxl_weights: Experiment with different styles. Examples:
    • Photorealistic: stable-diffusion-xl-base-1.0.
    • Anime-inspired: anime-art-diffusion-xl.

Guidance and Scaling:

  • guidance_scale: Higher values (20–50) enhance adherence to the prompt but may reduce creativity. Adjust based on desired style.
  • ip_adapter_scale and controlnet_conditioning_scale: Use mid-range values (0.5–0.8) for balanced effects. Extreme values may overfit or underfit the conditioning input.

Controlnet Conditioning:

  • pose_strength, canny_strength, and depth_strength:
    • Recommended range: 0.5–0.8 for subtle yet effective conditioning.
    • Use lower values (0.2–0.4) for minimal intervention.

Advanced Features for Instant ID Generate Avatar:

  • Scheduler:
    • For fast and smooth results, use DEISMultistepScheduler or DPMSolverMultistepScheduler.
    • For precision, try EulerDiscreteScheduler.
  • LCM Parameters:
    • lcm_num_inference_steps: Set between 5–8 for a balance between speed and quality.
    • lcm_guidance_scale: Values of 10–15 work best for controlled outputs.

Capabilities

High-Quality Output

The model excels in generating visually stunning images across diverse styles and resolutions.

Style Adaptability

Choose from a wide array of artistic weights to achieve desired aesthetic outcomes.

Precision Controls

Leverage pose, canny, and depth controls to craft outputs with fine detail and alignment.

What can I use for?

Creative Projects: Design unique illustrations, concept art, or storyboards.

Visualization: Generate detailed visuals for presentations or promotional material.

Experimentation: Explore artistic styles and techniques using pre-trained weights.

Things to be aware of

Generate a photorealistic portrait using stable-diffusion-xl-base-1.0 with fine-tuned controlnet settings.

Experiment with anime-inspired outputs using anime-art-diffusion-xl.

Combine pose control with a well-defined prompt to create dynamic, action-packed scenes.

Adjust guidance_scale and pose_strength to observe how the model interprets intricate instructions.

Limitations

Performance Variability: Results may vary significantly based on input prompt and style selection.

Pose Limitations: Poorly aligned or low-quality pose images can reduce output fidelity.

Complex Scenes: Highly intricate prompts may result in unexpected outputs or artifacts.

Controlnet Dependencies: Overuse of controlnets can sometimes overly constrain the creative potential of the model.

Output Format: PNG

Related AI Models

each-upscaler

Image Upscaler by Each AI

each-upscaler

Image to Image
each-faceswap-v1

Eachlabs Face Swap

each-faceswap-v1

Image to Image
flux-depth-pro

Flux Depth Pro

flux-depth-pro

Image to Image
flux-fill-pro

Flux Fill Pro

flux-fill-pro

Image to Image