Deep Learning

• Introduction
• Your first deep learning model
  • Create a code environment with the required packages
  • Create a Deep Learning analysis to solve a Prediction problem
  • Review the architecture of you Deep Learning model
  • Monitor the performance of your model during the training
• Model architecture
  • Build Keras model
    • input_shapes
    • n_classes
    • Layer dimensions
  • Compile the model
• Training
• Multiple inputs
  • Regular multi-feature inputs
  • Custom-processed single-feature inputs
• Using image features
  • Scoring images
• Using text features
• Runtime and GPU support
  • Code environment
  • Selection of GPU
  • Using multiple GPUs for training
• Advanced topics
  • Start with weights from a previously trained model
  • How is the model trained?
  • Advanced training mode
    • Build sequence
    • Fit model
      • Usage of metrics in Callbacks
• Troubleshooting
  • Using pre-trained models from Keras
  • Code environment lineage
  • TensorFlow session
  • ML API
  • Number of outputs in the model
  • Enforced code environment for Project