* fixed a bug about unzip

* added a test case for MnistModelLoader

* cleared warnings in the UnitTest project

* Upgrade NumSharp with signature.

* Testing gradients function: (#307)

- Concat
- LGamma
- Tanh
- Slice

* Adding new ctor for `Session` (#309)

Meant to be a shortcut to directly set some basic configuration.

E.g.

```
var sessionOptions = new SessionOptions();
sessionOptions.SetConfig(new ConfigProto
{
   LogDevicePlacement = true,
   InterOpParallelismThreads = 4,
   GpuOptions = new GPUOptions
   {
      AllowGrowth = true,
      PerProcessGpuMemoryFraction = 0.12
   }
});
var session = tf.Session(sessionOptions);
```

Adding `AbsGrad`

To be aligned with Tensorflow, introducing also `Sign` operation
and its gradient.

* fix strided_slice_grad type convention error. #307

* added a static load method for MnistModelLoader

* adjusted the assembly name for hub

* add Prediction function for Transfer Learning

* fixed the path combining issue

* Init TensorFlow.Text project. put all UnitTest and Example into one project.

* Predict image category in bootleneck file.

* Adding `IsNan` operation (#315)

Exposing also the `IsFinite` one.
Unit testing both operations.

* change bottleneck file to raw image as input for transfer learning.

* remove unnecessary projects (#316)

* implementing some additional methods (#317)

* got rid off the reference to the TF.NET

* updated project file for packaging

* removing unused namespaces

* implementing Tensor to scalar and scalar to tensor conversions

Change nest.flatten to generic.

2: Upgrade Microsoft.ML.TensorFlow.Redist to v1.14.

Adding the `BatchMatMul` operation.
Expanding `BasicOperations` to test the new operation.