
Deep Learning with Keras and Tensorflow
IBM · View on Coursera ↗
Or see the score while you browse.
Taught by Samaya Madhavan, Ricky Shi, Alex Aklson, Romeo Kienzler, Joseph Santarcangelo, Wojciech 'Victor' Fulmyk, JEREMY NILMEIER
Data ScienceComputer Science
About the course
Deep learning is revolutionizing many fields, including computer vision, natural language processing, and robotics. In addition, Keras, a high-level neural networks API written in Python, has become an essential part of TensorFlow, making deep learning accessible and straightforward.
Read more
Mastering these techniques will open many opportunities in research and industry. You will learn to create custom layers and models in Keras and integrate Keras with TensorFlow 2.x for enhanced functionality. You will develop advanced convolutional neural networks (CNNs) using Keras. You will also build transformer models for sequential data and time series using TensorFlow with Keras. The course also covers the principles of unsupervised learning in Keras and TensorFlow for model optimization and custom training loops. Finally, you will develop and train deep Q-networks (DQNs) with Keras for reinforcement learning tasks (an overview of Generative Modeling and Reinforcement Learning is provided). You will be able to practice the concepts learned using hands-on labs in each lesson. A culminating final project in the last module will provide you an opportunity to apply your knowledge to build a Classification Model using transfer learning. This course is suitable for all aspiring AI engineers who want to learn TensorFlow and Keras. It requires a working knowledge of Python programming and basic mathematical concepts such as gradients and matrices, as well as fundamentals of Deep Learning using Keras.
The dissent
The most-upvoted critical review (27 votes, 1.0★).
“It uses Tensorflow 1, which is outdated now - should be updated to use Tensorflow 2.”
Coursera's 4.4 comes from 1,050 ratings. Of those, 6.4% are 1–2★.
Starisk reads the reviews learners voted most helpful, and weights each by the square root of its votes: a review with 80 upvotes counts about five times one with 2. Unvoted reviews are not skipped out of laziness — by this model they carry no evidence that other students endorsed them.
That measures what the most-endorsed reviews say, not “true quality.” The rating distribution above is the whole population, so you can see the sample next to it.
Reviews read on 3 Nov 2026.Read them on Coursera ↗