Neural Network built from scratch using tensorflow-numpy

In the previous post, we have seen the key features of tensorflow-numpy and how easy it is to use alongside numpy. Also, we have compared the performance of both with a small sigmoid experiment.

Now we will build a small Neural network step by step using Tensorflow-Numpy and see how different it is from Numpy. This is to demonstrate how seamless it is to use tensorflow-numpy along with tensorflow operations.

Let’s start by importing all necessary packages.

Loading dataset

Building model

Gradients using tf.GradientTape

Wrapping with @tf.function

#tensorflow-numpy #numpy #tensorflow #neural-networks #scratch