Recurrent Neural Networks and TensorFlow Customization

This free online course in practical machine learning with TensorFlow will begin by introducing you to the concept of Recurrent Neural Networks (RNNs) and building machine learning models for sequential data. You will also learn about sequential dependencies of labels and the computation of Recurrent Neural Networks. You will be introduced to how loss is computed in the case of a simple RNN. You will also learn about the modified RNN architecture, vanishing gradient problem and Long Short Term Memory (LSTM).

The course then introduces the concept of using Recurrent Neural Networks to build models for time series data. You will also learn how to train an RNN model for time series, forecasting, model training and model prediction. Next, you will be introduced to machine learning models, text classification, overfitting, underfitting, regression and the architecture of a neural network model.

The course then explains TensorFlow Customization and how to extend the functionality of TensorFlow 2.0. You will also learn about the basic concepts behind tensors and how to use tensors with hardware accelerators. This course then explains in detail the main concepts behind Tensorflow Keras and how to build a simple model with tf.keras. You will also learn how to perform training, evaluation, prediction and defining custom callbacks.