Star us on Like our dataset format for AI? Give us a ⭐ on GitHub.

  • ActiveLoop
    • Solutions
      • Drones & Satellite Imagery

      • Store and process aerial imaging data with ease

      • Self-driving Cars and Robotics

      • Self-serve data pipelines from cameras, radar & lidar

      • Biomedical

      • Store and build data pipelines collected from biomedical images

      • IntelinAir
        Case Study

      • Learn more how Activeloop helped IntelinAir to process and generate datasets from millions acres of aerial imagery

      Company
      • About

      • Learn about the company, its members and its vision

      • Help Center

      • Get all of your questions answered by contacting our team

      • Careers

      • Build cool things that matter. From anywhere

      DocsResourcesPricing
  • Log in
Low AWS GPU usage? Achieve up to 95% GPU utilization in SageMaker with Hub
    • Back
      • Blog
      • News

    Low AWS GPU usage? Achieve up to 95% GPU utilization in SageMaker with Hub

    Low GPU utilization is a crucial problem of ML workflows. We present the seamless integration of AWS SageMaker and Hub that enables you to train models 2x faster and achieve 95% GPU utilization while streaming data from S3 to AWS SageMaker.
    • Davit Buniatyan

      Davit Buniatyan

      on Dec 3, 202113 min read

    • Share:

    • Previous
        • Blog

      • Data-centric AI enablers. Best data-centric MLOps tools in 2022

      • on Oct 18, 2021
    • Next
        • Tutorials

      • Weights & Biases and Hub - best practices for tasty classification models for computer vision

      • on May 19, 2021
  • Subscribe to our newsletter for more articles like this

  • Simple API for creating, storing, and collaborating on AI datasets at scale.

    • Solutions
      Drones & Satellite ImagerySelf-driving Cars and RoboticsBiomedical
    • Company
      AboutContact usCareersPrivacy PolicyTerms & Conditions
    • Resources
      BlogHumans in the Loop PodcastWhitepaper
  • Tensie

    Featured by