Generative AI - Enhancing Fine Tuning with More Training Data

Enhancing Fine-Tuning with More Training Data

We can add more training data to our fine-tuned model to make it work better. We will do the same things we did before, but this time we will use a bigger and more varied collection.

Step 1: Setup and Import Necessary Libraries

If you haven't already, install and load the tools you need.

# Install the OpenAI package
!pip install openai

# Import the library
import openai
import json

Step 2: Set Up API Key

Set up the key for the OpenAI API.

# Set up the API key
openai.api_key = 'your-api-key'

Step 3: Prepare the Extended Training Data

Put together a bigger collection with more examples.

# Extended training data
training_data = [
    {"prompt": "Translate English to French: 'Hello, how are you?'\n\n", "completion": "Bonjour, comment ça va?\n"},
    {"prompt": "Translate English to French: 'What is your name?'\n\n", "completion": "Comment t'appelles-tu?\n"},
    {"prompt": "Translate English to French: 'Good morning'\n\n", "completion": "Bonjour\n"},
    {"prompt": "Translate English to French: 'Good night'\n\n", "completion": "Bonne nuit\n"},
    {"prompt": "Translate English to French: 'How much is this?'\n\n", "completion": "Combien ça coûte?\n"},
    {"prompt": "Translate English to French: 'Where is the restroom?'\n\n", "completion": "Où sont les toilettes?\n"},
    {"prompt": "Translate English to French: 'I love you'\n\n", "completion": "Je t'aime\n"},
    {"prompt": "Translate English to French: 'See you later'\n\n", "completion": "À plus tard\n"},
    {"prompt": "Translate English to French: 'Thank you very much'\n\n", "completion": "Merci beaucoup\n"},
    {"prompt": "Translate English to French: 'You're welcome'\n\n", "completion": "De rien\n"}
]

# Save the training data to a file
with open('extended_training_data.jsonl', 'w') as f:
    for item in training_data:
        f.write(f"{json.dumps(item)}\n")

Step 4: Upload Training Data to OpenAI

Upload the extended training data file.

# Upload the training data file
response = openai.File.create(
  file=open('extended_training_data.jsonl', 'rb'),
  purpose='fine-tune'
)
training_file_id = response['id']

Step 5: Fine-tune the Model

Start the process of fine-tuning with the larger sample.

# Fine-tune the model
response = openai.FineTune.create(training_file=training_file_id, model="davinci")
fine_tune_id = response['id']

Step 6: Check Fine-tuning Status

Keep an eye on the process of fine-tuning.

# Check fine-tuning status
response = openai.FineTune.retrieve(id=fine_tune_id)
status = response['status']
print(f"Fine-tuning status: {status}")

Step 7: Use the Fine-tuned Model

Make text using the model that has been fine-tuned.

# Generate text using the fine-tuned model
response = openai.Completion.create(
  model=fine_tune_id,
  prompt="Translate English to French: 'Good evening'\n\n",
  max_tokens=50
)
print(response.choices[0].text.strip())

Full Code

# Install the OpenAI package
!pip install openai

# Import the library
import openai
import json

# Set up the API key
openai.api_key = 'your-api-key'

# Extended training data
training_data = [
    {"prompt": "Translate English to French: 'Hello, how are you?'\n\n", "completion": "Bonjour, comment ça va?\n"},
    {"prompt": "Translate English to French: 'What is your name?'\n\n", "completion": "Comment t'appelles-tu?\n"},
    {"prompt": "Translate English to French: 'Good morning'\n\n", "completion": "Bonjour\n"},
    {"prompt": "Translate English to French: 'Good night'\n\n", "completion": "Bonne nuit\n"},
    {"prompt": "Translate English to French: 'How much is this?'\n\n", "completion": "Combien ça coûte?\n"},
    {"prompt": "Translate English to French: 'Where is the restroom?'\n\n", "completion": "Où sont les toilettes?\n"},
    {"prompt": "Translate English to French: 'I love you'\n\n", "completion": "Je t'aime\n"},
    {"prompt": "Translate English to French: 'See you later'\n\n", "completion": "À plus tard\n"},
    {"prompt": "Translate English to French: 'Thank you very much'\n\n", "completion": "Merci beaucoup\n"},
    {"prompt": "Translate English to French: 'You're welcome'\n\n", "completion": "De rien\n"}
]

# Save the training data to a file
with open('extended_training_data.jsonl', 'w') as f:
    for item in training_data:
        f.write(f"{json.dumps(item)}\n")

# Upload the training data file
response = openai.File.create(
  file=open('extended_training_data.jsonl', 'rb'),
  purpose='fine-tune'
)
training_file_id = response['id']

# Fine-tune the model
response = openai.FineTune.create(training_file=training_file_id, model="davinci")
fine_tune_id = response['id']

# Check fine-tuning status
response = openai.FineTune.retrieve(id=fine_tune_id)
status = response['status']
print(f"Fine-tuning status: {status}")

# Generate text using the fine-tuned model
response = openai.Completion.create(
  model=fine_tune_id,
  prompt="Translate English to French: 'Good evening'\n\n",
  max_tokens=50
)
print(response.choices[0].text.strip())