2.1.0 is based on the Tensorflow 2.1.0 release.

Please check out our release note:
http://www.emgu.com/wiki/index.php/Emgu_TF_Version_History#Emgu.TF-2.1.0

Emgu TF Documentation can be found here:
http://www.emgu.com/wiki/files/emgutf/2.1.0/document/index.html

Emgu TF Lite Documentation can be found here:
http://www.emgu.com/wiki/files/emgutf/2.1.0/document-lite/index.html

1.14.0 is based on the Tensorflow 1.14.0 release

• For Android build, armeabi architecture is no longer supported. Android NDK removed armeabi architecture support.
• For iOS build, the full Tensorflow build is no longer supported. Tensorflow 1.14.0 recommends using Tensorflow Lite instead of full version of Tensorflow for iOS.
• The commercial release nuget packages have an additional "-CR" appended to it. e.g. The packaged id for the open source release is Emgu.TF and the package id for the similar commercial release has a package id of EMGU.TF-CR from our commercial repository. This will allow easier switching between the open source and commercial release.

Emgu TF v1.13.1 is based on the Tensorflow 1.13.1 release

• 1.13.1 is based on the Tensorflow 1.13.1 release
• For Android build, we use NDK r18b instead of r12b, armeabi is no longer supported.
• For the Xamarin demo project, we implemented image picking for Windows WPF platform.
• Bug fix for ImageIO.ReadTensorFromImageFile for the Windows platform when the image width is not a multiple of 4.
• Added the re-trained flower model in the Xamarin Forms demo.
• Improvements to error handling in the demo projects.
• Added TfInvoke.IsOperationSupported, TfInvoke.OpHasKernel

Emgu TF v1.12.0 is based on the Tensorflow 1.12.0 release

Emgu TF Lite release now also support Windows x64 platform.

Emgu TF v1.10.0 is based on the Tensorflow 1.10.0 release

The commercial version have added Emgu TF Lite package based on Tensorflow Lite that supports Android, iOS and Mac OS. Emgu TF Lite has a much smaller foot print and is best suited for Mobile development.