GPTpy is a Python boilerplate that utilizes OpenAI's GPT technology to generate human-like text in real-time. It includes presets for different output ranges, with formats tailored for education, standard usage, and creative prompts. Designed for self-driven work in a terminal environment, GPTpy is a versatile tool for personal projects and experimentation.
Make sure you use python3 instead of python to run Python 3 or create shell aliases for python and pip that always default to python3 and pip3
Create a Python virtual environment in the GPTpy directory before installing:
python -m venv env
Activate the virtual environment:
source env/bin/activate
Install the required Python packages:
pip install .
Add your API key from OpenAI
Create the .env file from your terminal in the project directory:
touch .env
Add the API key in the .env file:
OPENAI_API_KEY=...
Your project directory should now look like this:
gptpy
├─ env
├─ gptpy
│ ├─ __main__.py
│ ├─ api.py
│ ├─ interface.py
│ ├─ presets.py
│ ├─ system.py
├─ .env
├─ LICENSE.md
├─ README.md
├─ requirements.txt
├─ setup.py
Make sure you are inside the project directory then start GPTpy in your terminal:
python -m gptpy
Configure the GPT model hyperparameters manually in __main__.py:
(Note that user_prompt is added in the terminal input when the application is running)
#__main__.py
config(
model="text-davinci-003",
prompt=user_input,
temperature=0.5,
top_p=0.9,
max_tokens=500,
frequency_penality=0.5,
presence_penality=0
)Alternate from manual control to preset control by setting preset_control to True:
# __main__.py
preset_control = TrueIt is recommended to create shell aliases in your ~/.bashrc or ~/.zshrc file located in your home directory to always run Python 3:
alias python="python3"
alias pip="pip3"
If you always name your virtual environments as env you can create a shell alias to quickly activate the environment using the activate command:
alias activate="source env/bin/activate"
Here are some resourceful links for this project: