A Project named “TenserFlowOnSpark” is now open sourced by Yahoo.The Developer of “CafeeOnSpark” and model “Apache Spark” made it helpful for the developers to build learning models in Caffe.For more efficiency in Parallel Computing, an open source Framework Apache Spark was designed.
Even in a company like Netflix, Spark has created a remarkable difference by processing huge amount of user data thus offering a recommendation at scale.Without any use of rigorous Skill-Set TensorFlow and Caffe builds deep learning models.for machine learning.
By combining Caffe and Spark Yahoo created CaffeOnSpark which ultimately is a greater contribution to the machine learning community.For creating models even at a higher level of abstraction the open source libraries are existing for the developers.
To remove the gap between Spark and TensorFlow, the tools like TensorFrame and SparkNet was explored by Yahoo.This research resulted in TensorFlowOnSpark, the tool that enables developers to quickly modify their existing programs.
Skymind’s Deeplearning4j written for Java with Spark integrations is an open source framework which is in the sixth place in case of popularity. In Future Yahoo is planning for more improvement in the TensorFlowOnSpark.