TensorFlow w/R

tensorflow.md

Using TensorFlow in R

Chapter 1 - Your first "Hello World" neural network with TensorFlow in R

• Lesson 1.1 - Getting started with TensorFlow

  • A learning objective: Create tensors in R and use the building blocks of TensorFlow APIs

• Lesson 1.2 - Implementing graphs and loops in TensorFlow

  • A learning objective: Implement a multilayer neural network and evaluate it for predictions against a sample dataset

• Lesson 1.3 - Training a neural network with gradient descent

  • A learning objective: Use stochastic gradient descent and other training algorithms to improve neural network weights

Chapter 2 - TensorFlow in R for deep learning

• Lesson 2.1 - Implementing regression

  • A learning objective: Fit a deep neural network model to predict continuous values with mean squared error loss

• Lesson 2.2 - Implementing classification

  • A learning objective: Predict categorical values with a deep neural network using a cross-entropy loss function

• Lesson 2.3 - Improving model performance with feature engineering

  • A learning objective: Implement one-hot encoding, feature crosses, bucketization, and other custom features

• Lesson 2.4 - Training on out of memory data

  • A learning objective: Using tf.dataset and supporting libraries to train, validate, and test from sharded datasets.

Chapter 3 - Using Keras with TensorFlow in R

• Lesson 3.1 - Implementing customized layers in deep neural networks

  • A learning objective: Implement sequential and functional layers in Keras-based models.

• Lesson 3.2 - Using dropout for regularization

  • A learning objective: Extend Keras models with customized regularization layers and options.

• Lesson 3.3 - Getting started with hyperparameter tuning

  • A learning objective: Implement hyperparameter tuning for linear, logarithmic, and categorical hyperparameters.

• Lesson 3.4 - Setting up a machine learning pipeline

  • A learning objective: Create a pipeline to train, evaluate, and serve using multiple machine learning models

Chapter 4 - Using TensorFlow in R for image classification with convolutional neural networks

• Lesson 4.1 - Adding convolutional and maxpooling layers

  • A learning objective: Implement a convolutional neural network for image classification

• Lesson 4.2 - Adding batch normalization

  • A learning objective: Include batch normalization during deep neural network training

• Lesson 4.3 - Using image augmentation

  • A learning objective: Use tf.image library to augment image datasets with synthetic data

• Lesson 4.4 - Conclusion

  • A learning objective: Implement a deep neural network model for taxi fare prediction