Roboflow Changelog: June 2021
Each month, we publish a list of recent features and additions to the Roboflow suite of products. The previous month's update is here.
In May there has been a bevy of backend changes to support a huge user-facing release coming next month. The most noticeable updates this month were the launch of our OpenCV AI Kit integration and the addition of Classification dataset support in Roboflow Annotate.
Roboflow Organize
- Backend refactor to support upcoming changes to team sharing
- Preparation for a major user-facing release scheduled for next month 🤐
- User testing and iterating on the new UI we launched last month
- Updated dependency versions to pull in upstream bug fixes
Roboflow Train
- Improved autoscaling and concurrency for higher model training throughput
- Added additional memory to support larger datasets
- Added support for training classification models (currently in pre-release with enterprise customers; contact sales if you want access to the beta)
Roboflow Infer
- Added Luxonis OpenCV AI Kit devices as a deployment target
- Released new inference server Docker with support for more parameters from the remote inference API
- Unified class colors between annotation tool and prediction visualizations
- Updated to return predictions in order of confidence
- Refactored inference code to be shared across different model types
Roboflow Annotate
- Added support for editing classification labels
Other
- Added 2 new team members, Robert and Mohamed.
- Opened job listings for a Head of Marketing, a UX Designer, and a Business Development Representative.
- Added a Search feature to the blog.
- Added a page detailing Roboflow Deploy.
- 17 bug fixes.
- 1 new public dataset (Drone Gesture Control dataset).
- 13 blog posts.
- 14 YouTube videos.
- 3 Roboflow live demo webinars.
- Hosted a panel discussion for the incoming S21 Y Combinator batch.
- Fun: had a team onsite in Des Moines.
- Featured in/at: Rackspace SOLVE, Yahoo! Finance, InnovationIOWA, 1 Million Cups Iowa City, a Taiwanese computer science curriculum
- User projects: gesture-controlled drone, finding plastic in the ocean, synthetic data for package detection, detecting Leukemia in blood cells, recognizing defects in 3d printed parts, oil and gas risk avoidance.