View a Version

You can retrieve information about a Roboflow version using the Python SDK, REST API, and CLI.

Version ID

Each version generated in Roboflow has a sequential, numerical ID associated with it. You can see that through the app on your versions page or by listing your project versions in the API.

Python

To retrieve information about a project, use this command:

import roboflow

rf = roboflow.Roboflow(api_key=YOUR_API_KEY_HERE)

# get a project
project = rf.workspace().project("PROJECT_ID")

model = project.version("1").model

The model variable contains the following JSON values:

{
  "id": "mug-detector-eocwp/12",
  "name": "Mug Detector",
  "version": "12",
  "classes": null,
  "overlap": 30,
  "confidence": 40,
  "stroke": 1,
  "labels": false,
  "format": "json",
  "base_url": "https://detect.roboflow.com/"
}

REST API

CLI

Keypoint Detection Skeletons

Keypoint Detection project versions will contain a skeletons field which contains your version skeletons for each class. For example, a project with a class person might have the following skeleton:

{
...
"skeletons": {
    "person": {
        "vertices": [
            {
                "color": "#FF8000",
                "id": 0,
                "name": "nose",
                "x": 0.4546,
                "y": 0.18859999999999996
            },
            {
                "color": "#FF00FF",
                "id": 1,
                "name": "left_eye"
                "x": 0.478,
                "y": 0.1606
            },
            ...
        ],
        "edges": [
            {
                "color": "#00FFCE",
                "from": 13,
                "to": 15
            },
            ...
        ],
        "symmetries": [
            {
                "direction": "horizontal",
                "points": [1, 2]
            },
            ...
        ]
    }
}

Note, vertices should be accessed by index with inference predictions class_id. Then, vertex id can be used to reference edge (from/to) and symmetries (points).