Evaluate NSFW Content Detection with the fofr/nsfw-model-comparison Actions

23 Apr 2025
Evaluate NSFW Content Detection with the fofr/nsfw-model-comparison Actions

In today's digital landscape, understanding and managing explicit content is crucial for maintaining safe online environments. The fofr/nsfw-model-comparison provides a powerful Cognitive Action that helps developers compare various NSFW (Not Safe For Work) detection models against input images. This integration allows you to assess the effectiveness of different models, facilitating informed decisions for content moderation in applications.

Prerequisites

Before you start using the Cognitive Actions, ensure you have the following:

  • An API key for accessing the Cognitive Actions platform.
  • Basic knowledge of JSON and HTTP requests.
  • Familiarity with handling images via URIs.

Authentication for the Cognitive Actions typically involves passing your API key in the request headers.

Cognitive Actions Overview

Compare NSFW Models Against Inputs

This action evaluates multiple NSFW detection models to predict the level of explicit content in your input images. It is particularly useful for selecting the most accurate model based on comparative results.

  • Category: Content Moderation

Input

The input for this action requires a single field:

  • imageUri (required): A string representing the URI of the input image. It must be in a valid URI format and should point to the location of the image.

Example Input:

{
  "imageUri": "https://replicate.delivery/pbxt/LZ407zJWvzliNmeTeeb7VVlSo8IMdFanNlGEziaIHHxXvmcS/ComfyUI_00301_.png"
}

Output

The action returns a JSON object with the results of the model comparisons. The output contains:

  • falcon_is_safe: Boolean indicating whether the Falcon model considers the image safe.
  • compvis_is_safe: Boolean indicating whether the CompVis model considers the image safe.
  • falcon_time_taken: The time taken by the Falcon model to process the request (in seconds).
  • compvis_time_taken: The time taken by the CompVis model to process the request (in seconds).

Example Output:

{
  "falcon_is_safe": true,
  "compvis_is_safe": true,
  "falcon_time_taken": 0.11198115348815918,
  "compvis_time_taken": 0.26310253143310547
}

Conceptual Usage Example (Python)

Here’s how you might call the Cognitive Actions API to compare NSFW models 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 = "c612352d-3f81-478a-8979-6b01c149e932"  # Action ID for Compare NSFW Models Against Inputs

# Construct the input payload based on the action's requirements
payload = {
    "imageUri": "https://replicate.delivery/pbxt/LZ407zJWvzliNmeTeeb7VVlSo8IMdFanNlGEziaIHHxXvmcS/ComfyUI_00301_.png"
}

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 input payload is constructed based on the required schema, and the action ID is specified for the NSFW model comparison. The endpoint URL is illustrative, showcasing a generic request structure.

Conclusion

The fofr/nsfw-model-comparison Cognitive Action provides a straightforward and effective way to assess the performance of different NSFW detection models. By integrating this functionality into your applications, you can enhance content moderation efforts and ensure a safer user experience. For next steps, consider exploring additional Cognitive Actions that complement your existing moderation strategies or experiment with different images to further understand model behavior.

← Back to All Articles