Building Deep Learning Applications with Keras

Building Deep Learning Applications with Keras

1h 50mIntermediate2024-03-12

Authors

Isil Berkun

Isil Berkun

Data Scientist at Intel Corp.

Course details

Keras, a high-level neural networks API, is gaining popularity for its ease of use and versatility. With businesses rapidly moving towards AI solutions, you need an understanding of this valuable tool. In this course, data scientist Isil Berkun introduces you to Keras, highlights integrating it with TensorFlow and Theano backends, and offers practical insights into creating neural networks. Learn how to set up Keras, create neural networks, and use pre-trained models. Find out how to deploy models on platforms like Google Cloud. When you complete this course, you will be able to understand the Keras architecture, design and train deep learning models, and utilize them for real-world applications.

Skills covered

KerasTensorFlowNeural Networks and Deep LearningTelecommunicationsFull-Stack Web DevelopmentGoogleArtificial Intelligence (AI)Web DevelopmentNetwork and System AdministrationOpen SourceOne-Off

Concepts

Introduction

  • Reshaping the world with deep learning
  • Essential background and knowledge
  • How to use Codespaces and the exercise files

Understanding Keras

  • Understanding deep learning and Keras
  • Neuron as we know it
  • Exploring the TensorFlow and Theano backends
  • Distinction between Keras and TensorFlow

Setting up Keras

  • Keras installation with a TensorFlow backend on Windows

Getting Started with Keras Models

  • The Train-Test-Evaluate cycle
  • Introduction to the Keras Sequential API
  • Data pre-processing for training
  • Building a model using the Sequential API

Model Training and Performance Analysis

  • Training models
  • Model predictions and evaluation
  • Visualize results and save the model

Leveraging Pre-Trained Models in Keras

  • Exploring pre-trained models
  • Image recognition with the ResNet50 model

Tools for Visualization and Assessment

  • Exporting logs for Keras to TensorFlow
  • Monitoring training performance with TensorBoard
  • Visualizing computation graphs with TensorBoard

Conclusion

  • Next steps
40,000 Toman