Unlocking Creative Possibilities with pnickolas1/cog-experiments-v2 Cognitive Actions

21 Apr 2025
Unlocking Creative Possibilities with pnickolas1/cog-experiments-v2 Cognitive Actions

In the world of AI and machine learning, image generation has become a fascinating field, allowing developers to create stunning visuals from textual descriptions. The pnickolas1/cog-experiments-v2 API offers a powerful Cognitive Action that enables developers to generate images with inpainting and refinement capabilities. This action provides fine control over image attributes, making it ideal for applications in art generation, content creation, and more. By leveraging these pre-built actions, developers can save time and effort while enhancing their applications with advanced image processing capabilities.

Prerequisites

Before you start using the Cognitive Actions from the pnickolas1/cog-experiments-v2 API, ensure you have the following:

  • An API key for accessing the Cognitive Actions platform.
  • Basic knowledge of JSON and HTTP requests.
  • A development environment that can handle HTTP requests (e.g., Python, Postman).

Authentication typically involves passing the API key in the request headers to authorize access to the Cognitive Actions.

Cognitive Actions Overview

Generate Image with Inpainting and Refinement

The Generate Image with Inpainting and Refinement action is designed to produce images using img2img and inpainting modes. Developers can apply various parameters, including masks, prompts, and refinement styles, to create images that meet specific requirements.

Input

The input data for this action requires several fields, which are outlined in the schema below:

{
  "mask": "uri",
  "seed": 39536,
  "image": "https://replicate.delivery/pbxt/K3U2UvLjleEbsGJBZmlmSO8Sdreo5hwn3wV19Fo527beQvXA/IMG_0385.jpeg",
  "width": 1024,
  "height": 1024,
  "prompt": "line art of TOK, coloring book, black and white, white background, 4k, ",
  "loraScale": 0.6,
  "numOutputs": 1,
  "refineStyle": "no_refiner",
  "guidanceScale": 50,
  "highNoiseFrac": 0.8,
  "applyWatermark": true,
  "negativePrompt": "(low quality, worst quality:1.4), (bad anatomy), (inaccurate limb:1.2), \"bad composition, inaccurate eyes, extra digit, fewer digits, (extra arms:1.2), partial face, partial head, cropped head",
  "promptStrength": 1,
  "numInferenceSteps": 50,
  "schedulingAlgorithm": "DDIM"
}
  • mask (optional): URI of the mask for inpainting.
  • seed (optional): Random seed for reproducibility.
  • image (required): URI of the input image.
  • width (optional): Width of the output image in pixels (default: 1024).
  • height (optional): Height of the output image in pixels (default: 1024).
  • prompt (required): Text prompt guiding the image generation.
  • loraScale (optional): Scaling factor for trained models (0 to 1).
  • numOutputs (optional): Number of images to generate (1 to 4).
  • refineStyle (optional): Refinement technique applied (default: no_refiner).
  • guidanceScale (optional): Strength of classifier-free guidance (1 to 50).
  • highNoiseFrac (optional): Fraction of noise for expert refinement (0 to 1).
  • applyWatermark (optional): Whether to apply a watermark (default: true).
  • negativePrompt (optional): Traits to avoid in the generated image.
  • promptStrength (optional): Strength of the prompt (0 to 1).
  • numInferenceSteps (optional): Number of denoising steps (1 to 500).
  • schedulingAlgorithm (optional): Algorithm for scheduling inference steps (default: K_EULER).

Output

The action typically returns a URL pointing to the generated image, for example:

[
  "https://assets.cognitiveactions.com/invocations/415adc31-1a2e-4787-bed8-adf95e3d6ea1/fb6c86b5-31a3-4d0b-bc53-01ab9ce2b04c.png"
]

The output might contain variations based on the generated images, and error structures may also arise in case of issues during processing.

Conceptual Usage Example (Python)

Here's how a developer might call the Generate Image with Inpainting and Refinement action using Python:

import requests
import json

# Replace with your Cognitive Actions API key and endpoint
COGNITIVE_ACTIONS_API_KEY = "YOUR_COGNITIVE_ACTIONS_API_KEY"
COGNITIVE_ACTIONS_EXECUTE_URL = "https://api.cognitiveactions.com/actions/execute" # Hypothetical endpoint

action_id = "1621f7ed-bc99-4ca1-8867-6430872dc0ca"  # Action ID for Generate Image with Inpainting and Refinement

# Construct the input payload based on the action's requirements
payload = {
    "seed": 39536,
    "image": "https://replicate.delivery/pbxt/K3U2UvLjleEbsGJBZmlmSO8Sdreo5hwn3wV19Fo527beQvXA/IMG_0385.jpeg",
    "width": 1024,
    "height": 1024,
    "prompt": "line art of TOK, coloring book, black and white, white background, 4k, ",
    "loraScale": 0.6,
    "numOutputs": 1,
    "refineStyle": "no_refiner",
    "guidanceScale": 50,
    "highNoiseFrac": 0.8,
    "applyWatermark": True,
    "negativePrompt": "(low quality, worst quality:1.4), (bad anatomy), (inaccurate limb:1.2), \"bad composition, inaccurate eyes, extra digit, fewer digits, (extra arms:1.2), partial face, partial head, cropped head",
    "promptStrength": 1,
    "numInferenceSteps": 50,
    "schedulingAlgorithm": "DDIM"
}

headers = {
    "Authorization": f"Bearer {COGNITIVE_ACTIONS_API_KEY}",
    "Content-Type": "application/json"
}

try:
    response = requests.post(
        COGNITIVE_ACTIONS_EXECUTE_URL,
        headers=headers,
        json={"action_id": action_id, "inputs": payload}  # Hypothetical structure
    )
    response.raise_for_status()  # Raise an exception for bad status codes (4xx or 5xx)

    result = response.json()
    print("Action executed successfully:")
    print(json.dumps(result, indent=2))

except requests.exceptions.RequestException as e:
    print(f"Error executing action {action_id}: {e}")
    if e.response is not None:
        print(f"Response status: {e.response.status_code}")
        try:
            print(f"Response body: {e.response.json()}")
        except json.JSONDecodeError:
            print(f"Response body: {e.response.text}")

In this code snippet, replace YOUR_COGNITIVE_ACTIONS_API_KEY with your actual API key. The action ID and input payload are constructed according to the action's requirements. The endpoint URL and request structure are illustrative, focusing on how to manage the inputs and handle responses effectively.

Conclusion

The pnickolas1/cog-experiments-v2 Cognitive Actions provide a robust framework for developers looking to integrate advanced image generation capabilities into their applications. By using the Generate Image with Inpainting and Refinement action, you can unlock creative potential and enhance user engagement through captivating visuals. Experiment with different parameters to discover the full range of possibilities. Happy coding!

← Back to All Articles