What is DataGym?
DataGym is a comprehensive AI-powered training data and image annotation platform designed for professional teams working on computer vision projects. The platform combines AI-assisted labeling functions with robust project management capabilities to streamline the data preparation process.
The platform supports both image and video annotation, featuring specialized editors and AI-powered tools like Pre-Labeling that can detect up to 80 distinct object classes. Additionally, it offers seamless integration options through APIs, allowing users to stream data directly from various sources including local machines, public URLs, and AWS S3 buckets.
Features
- AI-Assisted Labeling: Accelerate annotation speed up to 10x faster
- Pre-Labeling: Automatic detection of up to 80 object classes
- Video Annotation: Specialized editor with adjustable playback speeds
- API Integration: Stream and export data seamlessly
- Multiple Import Sources: Support for local files, URLs, and S3 buckets
- Quality Control: Built-in systems for maintaining data accuracy
- Team Collaboration: Complete workbench for team management
Use Cases
- Computer vision model training
- Image dataset preparation
- Video annotation projects
- Machine learning data preparation
- Training data pipeline creation
- Object detection projects
FAQs
-
What types of files can I import into DataGym?
You can import image data from your local machine, any public image URL, or directly from an AWS cloud S3 bucket. -
How many object classes can the Pre-Labeling feature detect?
The Pre-Labeling feature can detect up to 80 distinct object classes on your images. -
What features are included in the free version?
The free Developer version includes 1 user, 3 projects, 400 images/month, 1GB storage, 100x AI assistance per month, API support, and external storage capabilities.
