Build: Prepare and upload a fine-tuning dataset

0/2 steps

✓📚 Read

✓👁 See

✓🎞 Animate

✓🎯 Test

5⚙️ Build

Build Steps

Progress

0/2

◆ Before you start

Prerequisites

Python 3.11+Required
python --version
openai>=1.30Required
pip install openai
OPENAI_API_KEYRequired
export OPENAI_API_KEY=sk-...
OpenAI paid accountRequired
Fine-tuning requires a funded account

📋

Step 1 of 2

Prepare and upload a fine-tuning dataset

Create a JSONL file with at least 10 chat examples in the required messages format, then upload it to the OpenAI files API.

Tasks

Create training.jsonl with 10+ chat examples in your target style

Validate each line parses as valid JSON with required fields

Upload the file using client.files.create(purpose="fine-tune")

Save the returned file ID for the next step

🎯 What You'll Learn

▸Understand the {messages: [{role, content}]} fine-tuning format
▸Validate JSONL structure before uploading
▸Use the Files API to upload training data

Code

Pythonprepare_finetune.py
1 import json
2 from openai import OpenAI
3  
4 client = OpenAI()
5  
6 # Create training examples
7 examples = [
8     {
9         "messages": [
10             {"role": "system", "content": "You are a concise technical assistant."},
11             {"role": "user", "content": "What is a REST API?"},
12             {"role": "assistant", "content": "A REST API is an interface that uses HTTP methods (GET, POST, PUT, DELETE) to perform CRUD operations on resources identified by URLs."},
13         ]
14     },
15     # Add 9+ more examples...
16 ]
17  
18 with open("training.jsonl", "w") as f:
19     for example in examples:
20         f.write(json.dumps(example) + "\n")
21  
22 # Upload to OpenAI
23 with open("training.jsonl", "rb") as f:
24     response = client.files.create(file=f, purpose="fine-tune")
25  
26 print(f"File ID: {response.id}")
27

💡 Pro Tips

▸Aim for 50-100 examples for meaningful style adaptation
▸Include edge cases and diverse phrasings in your training set

Discussion

1 import json 2 from openai import OpenAI 3 4 client = OpenAI() 5 6 # Create training examples 7 examples = [ 8 { 9 "messages": [ 10 {"role": "system", "content": "You are a concise technical assistant."}, 11 {"role": "user", "content": "What is a REST API?"}, 12 {"role": "assistant", "content": "A REST API is an interface that uses HTTP methods (GET, POST, PUT, DELETE) to perform CRUD operations on resources identified by URLs."}, 13 ] 14 }, 15 # Add 9+ more examples... 16 ] 17 18 with open("training.jsonl", "w") as f: 19 for example in examples: 20 f.write(json.dumps(example) + "\n") 21 22 # Upload to OpenAI 23 with open("training.jsonl", "rb") as f: 24 response = client.files.create(file=f, purpose="fine-tune") 25 26 print(f"File ID: {response.id}") 27

1	import json
2	from openai import OpenAI
3
4	client = OpenAI()
5
6	# Create training examples
7	examples = [
8	{
9	"messages": [
10	{"role": "system", "content": "You are a concise technical assistant."},
11	{"role": "user", "content": "What is a REST API?"},
12	{"role": "assistant", "content": "A REST API is an interface that uses HTTP methods (GET, POST, PUT, DELETE) to perform CRUD operations on resources identified by URLs."},
13	]
14	},
15	# Add 9+ more examples...
16	]
17
18	with open("training.jsonl", "w") as f:
19	for example in examples:
20	f.write(json.dumps(example) + "\n")
21
22	# Upload to OpenAI
23	with open("training.jsonl", "rb") as f:
24	response = client.files.create(file=f, purpose="fine-tune")
25
26	print(f"File ID: {response.id}")
27