DataGym Python API

Python API Wrapper for DataGym (https://www.datagym.ai) - the AI-assisted image annotation tool. You can easily access your Projects, Datasets, Images and labeled data from the DATAGYM backend through this package.

DATAGYM enables data scientists and machine learning experts to label images up to 10x faster. AI-assisted annotation tools reduce manual labeling effort, give you more time to finetune ML models and speed up your go to market of new products.

• Free software: BSD license
• Documentation: https://docs.datagym.ai/documentation/python-api/
• GitHub: https://github.com/datagym-ai/datagym-python

Requirements

• Python Version: Python 3.6, 3.7, 3.8, or later

• DataGym Account: If you haven't got a DataGym account yet, please SignUp.

• Initial Project: Setup your first Project and/or label some images in your Dummy Project

• API Token: In order to access your Dummy Project via our Python API you need to create an API Token

First Steps

Install the package via pip:

user@machine:~$ pip install datagym

Create your first DATAGYM API Client:

from datagym import Client

client = Client(api_key="<YOUR_API_KEY>")

Features

• Fetch Project, Datasets, Images
• Export labeled data from Projects
• Import labeled data into Projects
• Create Datasets
• Add Datasets to Projects
• Upload/Download images
• Upload of coco labels
• Upload of label configuration