Welcome to the Chat with LLM Models app! This interactive Streamlit application allows you to chat with powerful Large Language Models (LLMs) such as llama2 and deepseek-r1 in real-time. It’s designed to offer seamless conversations with AI, complete with model response streaming and performance tracking.
- Real-time Chat: Engage in live conversations with the selected LLM model.
- Model Selection: Choose between three cutting-edge models (llama2, deepseek-r1,gemma-2) via a simple sidebar.
- Response Streaming: Get responses streamed as they are generated, so you see the AI's thought process in real-time.
- Performance Tracking: Display response times for each interaction.
- Error Handling: Graceful error handling with helpful messages for troubleshooting.
- Logging: Every user input, model response, and error is logged for transparency and debugging.
Before running the app, ensure you have the following installed:
- Python 3.8+
- Streamlit
- llama_index (for interacting with the models)
- ollama (for model access)
To install the necessary dependencies, run the following:
pip install streamlit llama_index ollama
Clone or Download the Repository Clone the repo using Git:
bash Copy Edit git clone https://github.com/yourusername/chat-with-llm-models.git Or download the ZIP and extract it to your local machine.
Install Dependencies Install the necessary Python libraries by running:
bash Copy Edit pip install -r requirements.txt Run the App Start the Streamlit app with the following command:
bash Copy Edit streamlit run app.py Access the App Once the app is running, open your web browser and go to http://localhost:8501 to interact with the app.
Choose a Model Use the sidebar to select a model (llama2 or deepseek-r1) and start chatting!
Sidebar Model Selection: Choose your desired model from the dropdown menu. User Input: Enter your prompt in the chat input box. Live Response Streaming: As the model generates a response, you’ll see it stream live, letter by letter. Response Duration: The app displays how long it took the model to generate the response, so you can track performance.
User: "What's the capital of Japan?" Assistant: "The capital of Japan is Tokyo." Duration: The response time is displayed (e.g., Duration: 2.34 seconds).
📚 Logs & Error Handling Logs are automatically created for every user input, model response, and any errors that may occur. The app handles errors gracefully, providing useful messages if something goes wrong.
Try it out yourself! You can deploy this app to a platform like Streamlit Cloud or any other cloud hosting service to make it available for everyone.
We welcome contributions! If you want to improve the app or add features, feel free to fork the repository and submit a pull request. Please follow the standard contribution guidelines:
Fork the repo Create a new branch Make your changes Create a pull request 📜 License This project is licensed under the MIT License - see the LICENSE file for details.
🎉 Enjoy your chat with LLM models!