Roboflow Changelog: July 2021

Welcome to this month's installment of the Roboflow Changelog highlighting all the updates we've pushed in the past month. The update for last month can be found here.

In June we pushed our major revamp of the backend that supports collaboration between team members ("Workspaces"). It's now easier for teams to work together- even as individual team members join and leave - and share unified plan and billing information. Additionally, we extended our support for Classification datasets through to Roboflow Train and Roboflow Deploy.

Roboflow Organize

• Release of Workspaces to better support teams
• Front-end to support Classification in Roboflow Train and Deploy
• Completed migration to Tailwind
• Improved onboarding and conversion systems
• Created process for account deletion requests
• Integrated BigQuery for better (internal) querying

Roboflow Train

• Added support for Classification
• Expanded support for more GPUs and instance types
• Created a "priority" queue for paying customers' Train jobs

Roboflow Infer

• Released v2 of hosted inference (opens in a new tab) (refactored backend, simplified URL scheme, better parameter names, new workspace-aware API keys)
• Added support for Classification
• Added "offline mode" (air-gapped on-premise) to the Roboflow Inference Server (opens in a new tab) for Enterprise users.

Roboflow Annotate

• Improved Label Assist model's predictions

Other

• Moved into a new office
• Announced free Roboflow + OpenCV AI Kit course (opens in a new tab)
• Setup an AngelList page (follow us here) (opens in a new tab)
• Presented at the Y Combinator Startup Career Expo (opens in a new tab)
• Set a record of 1000 daily active users (opens in a new tab)
• Updated docs for inference (opens in a new tab), classification (opens in a new tab), and 3rd party training (opens in a new tab)
• 28 bug fixes
• 12 blog posts
• 9 YouTube videos (opens in a new tab)
• 4 Roboflow live demo webinars (opens in a new tab)
• Fun: founder meetup in Washington DC, visited customers in San Francisco
• Featured in/at: Import AI (opens in a new tab), Machine Learning Monthly (opens in a new tab), OpenCV Modelplace.ai, Super Startup Club (opens in a new tab)
• User projects: scorpion detection (opens in a new tab), fish identifier for scuba divers (opens in a new tab), product defect detector (opens in a new tab), orange finder (opens in a new tab), model evaluation (opens in a new tab), annotation speed run (opens in a new tab)