OpenAI GPT-3.5-Turbo Fine-Tuning Example

This is a Python script that demonstrates how to use OpenAI's GPT-3.5-Turbo fine-tuning API to train a model with your own data and then use the trained model for generating completions.

Requirements

• OpenAI Python package (v0.27.9+)
• An OpenAI API key

Setup

1. Install the OpenAI Python package:

   pip install openai

2. Set your OpenAI API key as an environment variable:

   export OPENAI_API_KEY='your-api-key'

3. Prepare your JSONL dataset and edit the FILENAME, SYSTEMPROMPT and USERPROMPT

4. Start the code

   python main.py

Usage

This script consists of the following steps:

1. prepare_data(): This function prepares the data for training. It opens a file containing the training data and creates a file object in OpenAI. It then checks the file status until it is processed.

2. train_model(): This function trains the model. It creates a fine-tuning job and checks the job status until it is finished.

3. use_model(): This function uses the trained model to generate a chat completion.

4. main(): This function runs the above functions in order.

To run the script, simply execute the Python file:

python main.py

Notes

• Make sure to replace example_data.jsonl with the path to your own training data file.
• The use_model() function does a test run with a system message "You are an AI." and a user message "test content goes here". You can change these messages to your own test.
• This script uses asyncio for handling the API calls. Make sure your Python environment supports asyncio.

Limitations

• The OpenAI API has rate limits. If you make too many requests in a short period of time, you may encounter rate limit errors.
• Fine-tuning a model can take a long time and consume a lot of resources. Make sure you have enough quota and resources before running the script.