GPT3.5 fine tuning dummy example

create_data.py

import os
import json
import typer

app = typer.Typer()

def process_scripts(input_prompts_file: str, scripts_folder: str, output_file: str):
    output_data = []

    # Read input prompts from the .jsonl file
    with open(input_prompts_file, 'r') as prompts_file:
        for line in prompts_file:
            prompt_data = json.loads(line)
            prompt_text = prompt_data['text'].strip()
            script_filename = prompt_data['meta']['script']
            script_path = os.path.join(scripts_folder, script_filename)

            if os.path.exists(script_path):
                with open(script_path, 'r') as script_file:
                    script_content = script_file.read().strip()

                message = {
                    "role": "system",
                    "content": "You are a system that generates Python scripts when given a prompt."
                }

                user_message = {
                    "role": "user",
                    "content": prompt_text
                }

                assistant_message = {
                    "role": "assistant",
                    "content": script_content
                }

                output_data.append({"messages": [message, user_message, assistant_message]})

    # Write the output data to the output file
    with open(output_file, 'w') as output:
        for entry in output_data:
            json.dump(entry, output)
            output.write('\n')

@app.command()
def main(
    input_prompts_file: str = typer.Argument(..., help="Path to the input prompts .jsonl file"),
    scripts_folder: str = typer.Argument(..., help="Path to the folder containing Python scripts"),
    output_file: str = typer.Argument(..., help="Path to the output .jsonl file")
):
    """
    Convert input prompts and scripts into output JSONL file.
    """
    process_scripts(input_prompts_file, scripts_folder, output_file)
    typer.echo("Conversion completed!")

if __name__ == "__main__":
    app()

hello_world.py

print("hello world!")

input_prompts.jsonl

{"text": "Write a Python script that runs hello world.", "meta": {"script": "hello_world.py"}}