Mastering Re-Ranking with the BGE-Reranker-Large Cognitive Actions

21 Apr 2025
Mastering Re-Ranking with the BGE-Reranker-Large Cognitive Actions

In the realm of natural language processing, effectively ranking responses based on relevance is crucial for building intelligent applications. The ninehills/bge-reranker-large spec offers a powerful Cognitive Action that allows developers to perform re-ranking using the BAAI/bge-reranker-large model. This action is designed to process paired inputs in JSON format, enabling applications to enhance their response selections based on the model's precision and capabilities.

Prerequisites

Before diving into the implementation, ensure you have the following:

  • An API key for the Cognitive Actions platform.
  • Basic knowledge of working with APIs and JSON.

Authentication typically involves passing your API key in the request headers, allowing you to securely access the Cognitive Actions.

Cognitive Actions Overview

Execute Re-Ranking with BAAI Model

The Execute Re-Ranking with BAAI Model action utilizes the BAAI/bge-reranker-large model with fp16 precision to perform re-ranking on a list of input pairs. This action is categorized under "other" and allows developers to enhance the relevance of their applications' responses.

Input

The input for this action requires a JSON object with a single key:

  • pairsJson: A string representing input pairs structured as JSON. This is a required field.

Example Input:

{
  "pairsJson": "[[\"a\", \"b\"], [\"c\", \"d\"]]"
}

Output

The output of this action typically consists of:

  • scores: An array of scores indicating the relevance of each pair.
  • use_fp16: A boolean indicating if fp16 precision was used.
  • model_name: The name of the model utilized for re-ranking.

Example Output:

{
  "scores": [
    0.64208984375,
    -0.82763671875
  ],
  "use_fp16": true,
  "model_name": "bge-reranker-large"
}

Conceptual Usage Example (Python)

Here’s how you might call this action in Python using a hypothetical Cognitive Actions execution endpoint:

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 = "ac4d1500-a0c8-4696-a9aa-173a1e25c976"  # Action ID for Execute Re-Ranking with BAAI Model

# Construct the input payload based on the action's requirements
payload = {
    "pairsJson": "[[\"a\", \"b\"], [\"c\", \"d\"]]"
}

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 values with your actual Cognitive Actions API key and endpoint. The payload variable contains the structured input JSON as specified by the action's requirements. The request is sent to the hypothetical execution URL, and the response is handled appropriately.

Conclusion

The BGE-Reranker-Large Cognitive Actions provide developers with a robust solution for enhancing the relevance of responses in their applications through effective re-ranking. By utilizing the BAAI model, you can ensure that your applications deliver higher quality results based on input pairs.

Consider exploring more use cases for these actions, such as integrating re-ranking in chatbots, search engines, or recommendation systems, to further improve user experience. Happy coding!

← Back to All Articles