Maybe two month later, I want to donate my scala keras style project build model base dependency tensorflow-java

[mullerhai]

HI :
tensorflow-java is Great Team, who make java & scala and other jvm language build model easily like python !! I feel appreciate about tensorflow-java all contributers. I have ask so many issue, and most the issues are quickly be answered by yours. I have build one framework make build model more easy ,I have write down more than 40+ tensorflow neural net algorithm Layers in my framework ,and users can also defined new layers for build model , they can use sequental style linear add layers build, also can build layer like model functional style , also can defined new model inherit super Model Module class, like defined Transformer Model, vgg , bert , It's inspire by keras-python , so I also named it scala-keras, I know also other coder has build similiar framework.

[pyrator]

That sounds great!

[karllessard]

Happy to help @mullerhai and thanks for the great comments! We also benefit having you looking in details at what is missing in our library so we can improve it in the future.

You can also join our community (SIG JVM) so we can look altogether at the issues you are facing and how we could help you to build your framework on top of tensorflow-java, we meet monthly on Google Meet, see the details here

[mullerhai]

Happy to help @mullerhai and thanks for the great comments! We also benefit having you looking in details at what is missing in our library so we can improve it in the future.

You can also join our community (SIG JVM) so we can look altogether at the issues you are facing and how we could help you to build your framework on top of tensorflow-java, we meet monthly on Google Meet, see the details here

Very fantastic ，I will clean and reformat my project code， then make the list about some tf function are need by the keras-scala framework,after I complete these ,I will join

[mullerhai]

Happy to help @mullerhai and thanks for the great comments! We also benefit having you looking in details at what is missing in our library so we can improve it in the future.

You can also join our community (SIG JVM) so we can look altogether at the issues you are facing and how we could help you to build your framework on top of tensorflow-java, we meet monthly on Google Meet, see the details here

I think if the Eager Session could support tf.varable() will be the most important things