Tensorflow is mysterious at the moment.
After initial training, I’d want Tensorflow to loop:
- feed a new set of observations to tf
- receive a set of predictions from tf
I’ve spent a few hours trying to find educational material - can anyone recommend a good place to go?
Hey,
For neural networks you’re probably better off using a wrapper on top of Tensorflow, unless you’re a machine learning expert/need the ability to do specific performance tuning. I’ve found Keras to be very easy to use for common network architectures such as dense, convolutional and fully connected layers.
Regarding educational material on the libraries, the Tensorflow website has several tutorials doesn’t it? Since so many people are bouncing off the walls for machine learning now, there is a lot of material on the internet you can find just by writing some precise/lucky google searches.
For the machine learning aspects of neural networks, which in my opinion are necessary for building and optimising an effective neural network, I’d highly recommend Michael Nielson’s online book, neuralnetworksanddeeplearning.com
The code.kx website has some tutorials on how to use embedpy, including integration with python libraries, such as Tensorflow or Keras.
Rob