Deep Learning (DL) · Advanced level

Advanced level · Module 4 of 6 · 22 guided hours

Deploying Deep Learning Models

Export, compress and serve deep learning models, including offline on phones and in low-connectivity settings.

Learning objectives

  • Export models to portable formats
  • Compress models with quantisation and pruning
  • Deploy models to servers and to phones for offline use

Topics

  • TorchScript and ONNX
  • quantisation
  • pruning and distillation
  • mobile runtimes
  • server inference and batching
  • monitoring deployed models

Lessons

  1. Lesson 1 Exporting Models

    Convert trained PyTorch models into portable formats for other runtimes.

    TorchScriptONNXruntimeexport verification

  2. Lesson 2 Compressing Models

    Make models smaller and faster with quantisation, pruning and distillation while keeping accuracy.

    quantisation (8-bit)accuracy droppruningknowledge distillation

  3. Lesson 3 Offline Deployment on Phones

    Run models directly on phones so they work without internet, and measure real performance.

    on-device inferencemobile runtimeslatency and memoryoffline-first design

  4. Lesson 4 Server Deployment and Monitoring

    Serve models from servers when needed, and monitor deep learning models after deployment.

    model serverbatchinginput monitoringfeedback collection

Practical activity

Export and quantise the leaf or pest model, measure size, speed and accuracy changes, and run it in an offline mobile or edge runtime.

Expected outcomes

  • Models that run efficiently where they are needed
  • Measured trade-offs from compression

Assessment

Deployment report with before-and-after measurements and a working offline demonstration.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.