The format required by OpenAI for fine-tuning GPT-4o for object detection.
Below, learn the structure of OpenAI GPT-4o Object Detection JSONL.
{
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "detect "},
{
"role": "user",
"content": [
{
"type": "image_url",
"image_url": {
"url": "URL"
}
}
]
},
{
"role": "assistant",
"content": " object ; object"
}
]
}
The loc tags correspond with the locations of objects in images. Individual annotations are separate by semi-colons, following the structure:
loc-tags class-name; loc-tags class-name;
With Roboflow supervision, an open source Python package with utilities for completing computer vision tasks, you can merge and split detections in OpenAI GPT-4o Object Detection JSONL. Read our dedicated guides to learn how to merge and split OpenAI GPT-4o Object Detection JSONL detections.
Below, see model architectures that require data in the OpenAI GPT-4o Object Detection JSONL format when training a new model.
On each page below, you can find links to our guides that show how to plot predictions from the model, and complete other common tasks like detecting small objects with the model.