FiftyOne Enterprise 2.14.0 Released with Auto Labeling and Native Kubernetes Orchestration
Blog post from Voxel51
FiftyOne Enterprise 2.14.0 introduces significant enhancements, including an Auto Labeling Panel for interactive human-in-the-loop annotation using foundation models and a Native Kubernetes Orchestrator that enables ephemeral compute at scale, optimizing resource use and efficiency. The release also features improvements in user experience, such as increased plugin upload size, enhanced data quality charts, and dynamic operator inputs. Security and resilience are bolstered through clean task terminations, dependency upgrades, and improved error handling. The new version aims to streamline workflows, increase productivity, and integrate seamlessly with existing infrastructure, positioning FiftyOne Enterprise as a unified platform for data, models, and teams to enhance visual AI performance.