2. Fine-tuning the model

To train the model, we retrieve our training file and create a fine-tuning job. We begin by creating an OpenAI client:

from openai import OpenAI
import jsonlines
client = OpenAI()

Then we use the file we generated to create another training file that is uploaded to OpenAI:

# Uploading the training file
result_file = client.files.create(
  file=open("QA_prompts_and_completions.json", "rb"),
  purpose="fine-tune"
)

We print the file information for the dataset we are going to use for fine-tuning:

print(result_file)
param_training_file_name = result_file.id
print(param_training_file_name)

We now create and display the fine-tuning job:

# Creating the fine-tuning job
 
ft_job = client.fine_tuning.jobs.create(
  training_file=param_training_file_name,
  model="gpt-4o-mini-2024-07-18"
)
# Printing the fine-tuning job
print(ft_job)

The output first provides the name of the file, its purpose...