List view
This release should bring significant performance and stability improvements, especially for the MacOSX users. Major highlights: - now tested against Spark 2.1 - uses the official (experimental) Java API for TensorFlow published by Google - groupBy, and the k-means example, is fixed
No due date•8/14 issues closedThis release will focus on: - convenience features for tensorflow users - mllib support may slip into this release - more performance improvements
No due date•1/5 issues closedThis release will focus on: - stabilizing the API (especially the data storage of high-dimension tensors) - experimental integration with catalyst - integration with mllib (potentially) Target spark version is 2.0+ Target tensorflow version is 0.8 or above.
No due date•5/11 issues closed