Enhance Image Generation with lucataco/ssd-lora-inference Cognitive Actions

23 Apr 2025
Enhance Image Generation with lucataco/ssd-lora-inference Cognitive Actions

In the world of AI-powered image generation, the lucataco/ssd-lora-inference API offers a remarkable set of capabilities through its Cognitive Actions. These actions harness the power of the SSD-1B model with Low-Rank Adaptation (LoRAs) to create and modify images in innovative ways. By leveraging these pre-built actions, developers can save time and effort while customizing image outputs to meet specific needs.

Prerequisites

To get started with the Cognitive Actions, you'll need an API key for the lucataco Cognitive Actions platform. Authentication typically involves passing your API key in the headers of your requests. Ensure you have the endpoint URL ready for executing the actions.

Cognitive Actions Overview

Run Inference on SSD-1B LoRAs

The Run Inference on SSD-1B LoRAs action allows you to execute inference using the SSD-1B model. This action can generate or modify images, offering extensive customization through various parameters such as prompts, masks, and image specifications.

Category: Image Generation

Input: The input schema requires the following fields, with some optional parameters for enhanced control:

  • loraUrl (required): URL to load the LoRA model from.
  • prompt (default: "A photo of TOK"): Text prompt to guide image generation.
  • mask (optional): URI of the input mask for inpaint mode.
  • image (optional): URI of the input image for img2img or inpaint mode.
  • seed (optional): Random seed for generating image variations.
  • width (default: 1024): Width of the output image in pixels.
  • height (default: 1024): Height of the output image in pixels.
  • loraScale (default: 0.6): Additive scale factor for the LoRA model.
  • scheduler (default: "K_EULER"): Method for scheduling the denoising process.
  • guidanceScale (default: 7.5): Scale for classifier-free guidance.
  • applyWatermark (default: true): Applies a watermark to the generated image.
  • negativePrompt (optional): Textual input to specify what should not appear in the image.
  • promptStrength (default: 0.8): Degree of adherence to the prompt.
  • numberOfOutputs (default: 1): Number of images to generate (1-4).
  • numberOfInferenceSteps (default: 50): Number of denoising steps used.

Example Input:

{
  "seed": 37543,
  "width": 1024,
  "height": 1024,
  "prompt": "A photo of TOK",
  "loraUrl": "https://replicate.delivery/pbxt/u5hevTlT560fI0D1TYxwozJk3gEJHAjVCubvbzngsaeoIIqjA/trained_model.tar",
  "loraScale": 0.6,
  "scheduler": "K_EULER",
  "guidanceScale": 7.5,
  "applyWatermark": true,
  "negativePrompt": "",
  "promptStrength": 0.8,
  "numberOfOutputs": 1,
  "numberOfInferenceSteps": 25
}

Output: Upon successful execution, the action typically returns an array of URLs pointing to the generated images.

Example Output:

[
  "https://assets.cognitiveactions.com/invocations/59e1508d-1bc5-461e-99ed-0d3cd917a965/6c2d8645-3dab-4815-830d-88372806eb36.png"
]

Conceptual Usage Example (Python): Here’s how you might call this action in 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 = "7127fa05-0f91-422a-98d3-7c30f35ad2ac"  # Action ID for Run Inference on SSD-1B LoRAs

# Construct the input payload based on the action's requirements
payload = {
    "seed": 37543,
    "width": 1024,
    "height": 1024,
    "prompt": "A photo of TOK",
    "loraUrl": "https://replicate.delivery/pbxt/u5hevTlT560fI0D1TYxwozJk3gEJHAjVCubvbzngsaeoIIqjA/trained_model.tar",
    "loraScale": 0.6,
    "scheduler": "K_EULER",
    "guidanceScale": 7.5,
    "applyWatermark": True,
    "negativePrompt": "",
    "promptStrength": 0.8,
    "numberOfOutputs": 1,
    "numberOfInferenceSteps": 25
}

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 the placeholder API key and endpoint with actual values. The action_id corresponds to the action you are invoking, and the payload is structured according to the input schema described.

Conclusion

The lucataco/ssd-lora-inference Cognitive Actions provide developers with powerful tools for image generation and modification. By integrating these actions into your applications, you can easily create customized images and enhance user experiences. Explore the different parameters available and experiment with various configurations to unlock the full potential of the SSD-1B model with LoRAs. Happy coding!

← Back to All Articles