Roboflow Changelog: March 2021

In March 2021, Roboflow focused on reinforcing its foundational infrastructure, fixing bugs, and preparing for future developments following a period of rapid feature expansion. Key updates included enhancements in Roboflow Organize, such as the introduction of Upload Batches and an Unannotated Queue to streamline labeling workflows, along with improved team collaboration features and caching mechanisms. Roboflow Train saw the groundwork laid for on-device deployments and the soft launch of roboflow.js for browser-based inference. Roboflow Infer added new preprocessing options and enhanced prediction visualization, while Roboflow Annotate boosted speed and reliability through WebGL context reuse, reduced CPU usage, and introduced a Label Assist onboarding flow. The company also announced key partnerships with OpenCV and collaborations with Paperspace and Streamlit, introduced new industry pages, and released additional content such as blog posts and YouTube videos. Highlights included welcoming a new Head of Customer Development and showcasing user projects like a cattle counter and social distancing drone.