chatter

Chatter User

Overview

chatter-user.py is a Python script that facilitates conversations between two selected Large Language Models (LLMs). The script allows users to configure and select models, send requests, and log the conversations.

Features

• Configuration: Supports multiple LLMs with specific configurations, including API keys for external models.
• Dynamic Model Selection: Users can select two models for interaction.
• User Interaction: Prompts for task input, system prompts, and number of turns for the conversation.
• API Requests: Sends requests to the specified LLM endpoints and processes responses.
• Conversation Logging: Saves the conversation to a timestamped file.
• Validation: Ensures valid inputs for the number of turns and model selection.

Usage

1. Set Up Environment Variables:

   • Ensure the necessary API keys are set in your environment variables:

     export EXTERNALMODEL_API_KEY=your_externalmodel_api_key

2. Clone the Repository:

   git clone <repository-url>
   cd <repository-directory>

3. Run the Script:

   python chatter-user.py

4. Interact with the Script:

   • Follow the prompts to select models, enter the task, system prompts, and number of turns.
   • The script will facilitate the conversation between the selected models and log the conversation to a file.

Example

$ python chatter-user.py
Available models: 1. Mixtral, 2. Gwen, 3. ExternalModel
Enter two model numbers separated by a comma (e.g., 1,2): 1,3
Enter the initial task/message for the LLMs: Discuss the impact of AI on society
Enter number of turns (2-100, default 10): 5
Enter system prompt base for Model 1: You are Mixtral. Discuss the topic with {other_model}.
Enter system prompt base for Model 3: You are ExternalModel. Discuss the topic with {other_model}.
Mixtral: AI has the potential to greatly impact society in various ways...
ExternalModel: Indeed, AI could revolutionize industries and improve efficiency...
...
Conversation saved to Discuss_the_imp_20240515_121045.txt
Token usage: {'total_tokens': 1200}

Configuration

• llm_configs: Configuration dictionary for the LLMs, including names, base URLs, ports, and API keys.

Future Expansion Plans

1. Enhanced Error Handling: Improve error messages and provide options for recovery.
2. Customizable Conversations: Allow users to customize more conversation parameters.
3. Additional Models: Integrate more LLMs for greater variety in interactions.
4. Advanced Metrics: Implement detailed metrics for tracking performance and usage.

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss your ideas.

License

This project is licensed under the MIT License.