Course

Customising your models with TensorFlow 2

Imperial College London

Welcome to the course on Customising your models with TensorFlow 2. Deepen your knowledge and skills with TensorFlow to develop fully customised deep learning models and workflows for any application.

This course will cover:

  • Lower level APIs in TensorFlow to develop complex model architectures
  • Fully customised layers and a flexible data workflow
  • Expansion of knowledge of the TensorFlow APIs to include sequence models
  • Practical, hands-on coding tutorials guided by a graduate teaching assistant
  • Automatically graded programming assignments for skill consolidation
  • A Capstone Project to develop a custom neural translation model from scratch

Prerequisite knowledge required includes proficiency in Python, general machine learning concepts, and a working knowledge of deep learning.

Certificate Available ✔

Get Started / More Info
Customising your models with TensorFlow 2
Course Modules

Deepen your knowledge with modules covering the Keras functional API, Data Pipeline, Sequence Modelling, Model subclassing and custom training loops, and a Capstone Project.

The Keras functional API

Welcome to Customising your Models with TensorFlow 2. This module covers the Keras functional API, multiple inputs and outputs, variables, tensors, freezing layers, and device placement. It also includes coding tutorials, knowledge checks, and additional readings to reinforce learning.

Data Pipeline

Welcome to week 2 - Data Pipeline. In this module, you will learn about Keras datasets, dataset generators, image data augmentation, the Dataset class, training with datasets, and TensorFlow Datasets. The module also includes coding tutorials, knowledge checks, and additional readings for comprehensive understanding.

Sequence Modelling

Welcome to week 3 - Sequence Modelling. This module covers preprocessing sequence data, the IMDB dataset, padding and masking sequence data, the Embedding layer, recurrent neural network layers, stacked RNNs, Bidirectional wrapper, and a language model for the Shakespeare dataset. It includes coding tutorials and readings for in-depth learning.

Model subclassing and custom training loops

Welcome to week 4 - Model subclassing and custom training loops. This module delves into model subclassing, custom layers, automatic differentiation, custom training loops, the tf.function decorator, and a residual network. Coding tutorials and readings supplement the learning experience.

Capstone Project

Welcome to the Capstone Project. This module is devoted to the Capstone Project, where you will apply all the concepts learned throughout the course to develop a custom neural translation model from scratch. The module also includes a post-course survey for feedback.

More Machine Learning Courses

Machine Learning with TensorFlow on Google Cloud en Français

Google Cloud

Machine Learning with TensorFlow on Google Cloud en Français offers hands-on training in ML using Google Cloud Platform. Learn to create, train, and deploy ML models,...

Intro to TensorFlow 日本語版

Google Cloud

Intro to TensorFlow 日本語版 provides hands-on training in building, training, and deploying machine learning models using TensorFlow 2.x and Keras.

Principal Component Analysis with NumPy

Coursera Project Network

Principal Component Analysis with NumPy is a 2-hour project-based course where you'll learn to implement PCA from scratch using Python, conduct EDA, and create data...

Generative AI Fundamentals

IBM

Generative AI Fundamentals is a comprehensive specialization offering a deeper understanding of generative AI, from foundational concepts to practical applications,...