migueldeicaza
Introduction text, what does train in python mean...
"You can prototype using TensorFlow or Keras in Python, then save your graphs or trained models and then load the result in .NET with TensorFlowSharp and feed your own data to train or run."
Can anyone help me understand it's kind of ambiguous... Especially since it suggest to save the "trained models" and then load them into .net... and then train them more ? This would not make sense.. why train and train again..
There are other ways to save tensorflow libraries and load them in C# and run them in recall mode or evaluate mode from there. If this library can actually setup , train , valuate and save models using C# it would really be nice to be made clear on the first page.. and if it cannot train that sentence is also confusing.
I am sorry for having to ask this. Yes I could spend fifteen minutes reading and get the answer. However this is not only for my understanding but for everyone else who comes across this library and this is probably one of the first paragraph that they read and if misunderstood might make them close the browser tab... or waste a lot of cumulative time trying to find a way to train the models in c# .
It means that you can build the graph and train the model in Tensorflow or Keras, then export the trained model to .pb file. Then use tensorflowsharp to load this .pb file and make inference.
Thank you for the quick answer. for full support including training , it would be better to use and assist with tensorflow.net. I am not affiliated to that project, but perhaps in the future.