kkk324 / chat-with-gpt4-streamlit

Chat with your GPT-4 model using OpenAI API via Streamlit web app

chat-with-gpt4-streamlit

Chat with your GPT-4 model using OpenAI API via Streamlit web app using latest OAI and streamlit packages

Main reference for Streamlit related code:

• https://docs.streamlit.io/knowledge-base/tutorials/build-conversational-apps

Blog article describing this app

Starting the app

1. Set up OpenAI API key as a streamlit secret

• Create .streamlit/secrets.toml file in the project/current directory and add the following lines to it:

OPENAI_API_KEY = "YOUR_API_KEY" NOTE: You need to procure your API key from the OpenAI website by making the required payment. You could start small with $5.

2. Set up virtual environment using venv in Unix/macOS:

• python -m venv .venv
• source .venv/bin/activate
• pip install -r requirements.txt

3. Start the app by following below steps:

• Activate the virtual environment: source .venv/bin/activate (if not already activated)

• Then, use Makefile command: make run or directly start app using streamlit run chat_with_gpt4_streamlit/main.py

• Navigate to http://localhost:8501 where streamlit runs by default

• Sample run of the app:

4. More features of the chat app:

• Streaming generation of the response - No waiting!
• Export the current conversation to save on API calls - outputs are written in the project directory inside a folder called exports
  • A CSV file is created with the time stamp of the local time when the export button was clicked
  • Export button:
  • Exported content as a CSV file:
• Get total tokens used using the tiktoken library
  • See total tokens used in the current conversation:

5. Advanced configurations:

• Adjust the name of the OpenAI text generation model used by changing the OAI_MODEL parameter in the configs.py file List of all possible models from OpenAI here and here
• Adjust the location where the chat exports are saved by changing the EXPORT_DIR parameter in the configs.py file. This path is with respect to the project root directory.