Course Code: bspdltk
Duration: 21 hours
Course Outline:
  • Install TensorFlow and tf.keras
    • What Are Keras and tf.keras?
    • Install TensorFlow using conda
  • Deep Learning Model Life-Cycle
    • The 5-Step Model Life-Cycle
    • Sequential Model API (Simple)
    • Functional Model API (Advanced)
  • Deep Learning Models
    • Multilayer Perceptron
      • Exercise - Build an MLP to solve regression task using TF
    • Convolutional Neural Networks
      • Exercise -  Image Recognition
      • Exercise - Object Detection using YOLOv3
    • Recurrent Neural Networks
      • Exercise - Generate Text
      • Exercise - Speech Recognition
  • Advanced Model Features
    • Visualize a Deep Learning Model
    • Plot Model Learning Curves
    • Save and Load Your Model
    • Training and Testing visualization using TensorBoard
  • Tuning Model Performance
    • Hyperparameter tuning frameworks
    • Accelerate Training With Batch Normalization
    • Halt Training at the Right Time With Early Stopping
    • Exercise - Use hyperparameter tuning framework to find optimal params
  • Deploy a model on ML Azure / AWS Cloud
    • Train the model on cloud
    • Create a REST API endpoint to use the model in an application

Note - To be able to work on ML Azure / AWS Cloud, participants will need an Azure account or AWS account. First time users can get free credits upto 500 GBP from Microsoft or AWS.