GraphLab Create: large-scale machine learning platform for graph, structured, and text data

GraphLab Create 1.0 brings large-scale machine learning capabilities to enterprises, and is the first to handle graph, structured, and text data in one platform.

By Gregory Piatetsky, Jul 15, 2014.

Seattle-based startup GraphLab announced today
GraphLab GraphLab Create 1.0, a new platform to bring large-scale machine learning capabilities to enterprise. GraphLab Create already has hundreds of beta-users.

GraphLab key advantages are performance and scaling. GraphLab Create has optimized out-of-core scaling enabling iteration of 1TB of data on a single machine, and can easily scale to distributed compute in the cloud for production environments that have high data volume and velocity.

News highlights:
  • GraphLab is the first to apply machine learning for 3 main types of data: graph, tabular and text data - in one platform.
  • The GraphLab engine is reported to be 100-10,000 times faster at analytics operations and model training over industry alternatives.
  • Users can prototype in a single machine and move to production with the same code
  • GraphLab data visualization capabilities simplify data exploration
  • GraphLab is integrated with Hadoop (HDFS), is Cloudera certified, and is available as part of Pivotal HD Hadoop distribution platform

GraphLab Create 1.0 will be available on July 21, 2014.

Additional Resources: