OpenAIError
openai.OpenAIError (fine-tuning job failure)
Stack trace
openai.OpenAIError: Fine-tuning job failed: insufficient training examples provided. Minimum required examples not met.
Why it happens
OpenAI fine-tuning requires a minimum number of labeled training examples to create a reliable model. If the uploaded dataset contains fewer examples than the minimum threshold, the fine-tuning job is rejected with this error.
Detection
Check the number of training examples in your fine-tuning dataset before submitting the job; log the dataset size and validate it meets OpenAI's minimum requirements.
Causes & fixes
Training dataset contains fewer examples than OpenAI's minimum required for fine-tuning.
Add more labeled examples to your training dataset to meet or exceed the minimum example count specified by OpenAI.
Dataset file uploaded to OpenAI is empty or corrupted, resulting in zero usable examples.
Verify the dataset file content and format before upload; ensure it is not empty and follows OpenAI's JSONL fine-tuning format.
Incorrect file uploaded (e.g., validation or test set instead of training set).
Confirm you are uploading the correct training dataset file for fine-tuning, not validation or test data.
Code: broken vs fixed
from openai import OpenAI
import os
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
# This dataset has too few examples, causing the error
training_file_id = "file-abc123"
response = client.fine_tunes.create(training_file=training_file_id) # Triggers insufficient examples error
print(response)from openai import OpenAI
import os
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
# Use a training file with sufficient examples
training_file_id = "file-valid123"
response = client.fine_tunes.create(training_file=training_file_id) # Fixed: dataset has enough examples
print(response)Workaround
If you cannot immediately add more examples, consider using a smaller base model or pre-trained model without fine-tuning until you gather sufficient training data.
Prevention
Always validate your training dataset size and format before uploading to OpenAI; automate checks to ensure minimum example counts are met to avoid job failures.