Generate Engaging Chat Responses with Qwen 2 Cognitive Actions

23 Apr 2025
Generate Engaging Chat Responses with Qwen 2 Cognitive Actions

Integrating advanced language models into applications has become a game-changer for developers looking to enhance user interactions. The Qwen 2 model, a 1.5 billion parameter language model from Alibaba Cloud, is designed specifically for generating chat completions. With its fine-tuning capabilities, it supports various tasks, including language understanding, generation, and even coding. This blog post will guide you through leveraging the Qwen 2 Cognitive Actions to create engaging and contextually relevant chat experiences.

Prerequisites

Before diving into the integration of Cognitive Actions, ensure you have:

  • An API key for accessing the Cognitive Actions platform.
  • Basic knowledge of JSON and RESTful API concepts.
  • A programming environment set up for making HTTP requests.

Authentication is typically done by passing your API key in the headers of your requests, allowing you to securely access the Cognitive Actions functionality.

Cognitive Actions Overview

Generate Chat Completions with Qwen 2

Description: This action utilizes the Qwen 2 model to generate chat completions based on user-defined prompts. It excels in various language tasks, providing responses that can be tailored to specific contexts.

Category: Chatbots

Input

The input for this action is structured as follows:

{
  "seed": 12345, 
  "prompt": "Tell me a joke about only having 1.5 billion parameters",
  "modelType": "Qwen2-1.5B-Instruct",
  "topKTokens": 1,
  "temperature": 1,
  "systemPrompt": "You are a funny and helpful assistant.",
  "topPPercentage": 1,
  "repetitionPenalty": 1,
  "maxGeneratedTokens": 512
}

Key Properties:

  • seed (integer, optional): A random seed for generating results.
  • prompt (string, required): The initial input to guide the model's response.
  • modelType (string, required): Specifies the model variant to use (default is Qwen2-1.5B-Instruct).
  • topKTokens (integer, optional): Controls the diversity of generated tokens.
  • temperature (number, optional): Modulates randomness in the output.
  • systemPrompt (string, optional): Sets the overarching behavior of the assistant.
  • topPPercentage (number, optional): Adjusts the sampling strategy for token generation.
  • repetitionPenalty (number, optional): Penalizes repeated phrases in the output.
  • maxGeneratedTokens (integer, optional): Limits the number of tokens generated in the response.

Output

When invoked, this action typically returns a structured array of strings representing the generated tokens. For example, if the input prompt is a joke request, the output might look like this:

[
  "",
  "Why ",
  "did ",
  "the ",
  "neural ",
  "network ",
  "have ",
  "only ",
  "",
  "",
  "",
  1.5,
  "billion ",
  "",
  "parameters? ",
  "Because ",
  "it ",
  "was ",
  "a ",
  "tiny ",
  "little ",
  "",
  "",
  "network!"
]

This output can be processed to form coherent responses suitable for chat interactions.

Conceptual Usage Example (Python)

Here’s how you might structure a Python script to call this action:

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 = "2b70e225-e4dc-4c5a-b3e3-4fe7d2b44e29"  # Action ID for Generate Chat Completions with Qwen 2

# Construct the input payload based on the action's requirements
payload = {
    "prompt": "Tell me a joke about only having 1.5 billion parameters",
    "modelType": "Qwen2-1.5B-Instruct",
    "topKTokens": 1,
    "temperature": 1,
    "systemPrompt": "You are a funny and helpful assistant.",
    "topPPercentage": 1,
    "repetitionPenalty": 1,
    "maxGeneratedTokens": 512
}

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, you replace the placeholders with your actual API key and endpoint. The input payload is structured according to the specifications, and the output from the API will give you the generated tokens, which you can then format for user-facing applications.

Conclusion

Utilizing the Qwen 2 Cognitive Actions can significantly enhance your application's ability to engage users with intelligent chat responses. The flexibility of input parameters allows you to tailor interactions to suit various contexts, making it a powerful tool in any developer's arsenal. Explore the possibilities and consider integrating these actions into your next project to create a more interactive and enjoyable user experience!

← Back to All Articles