Add support for the Spark-Tfrecord library

Description

LinkedIn's new Spark-Tfrecord library is an upgrade on the existing spark-tensorflow connector.
This Jira adds support for it in both Java/Scala and Python.
https://github.com/logicalclocks/spark-tfrecord

Also hops-util-py added support for tf2. For example replacing tf.FixedLenFeature with tf.io.FixedLenFeature

Assignee

Davit Bzhalava

Reporter

Jim Dowling

Labels

None

Sprint

None

Fix versions

Priority

Medium
Configure