AI: Local AI & Model Types
Model compression: quantise, distil, prune | Sri AI
Advanced
2 views
Course overview
Making large models small and fast enough for the hardware Sri Lankan institutions can actually buy.
Level: Advanced · Mode: Evenings, online
Who this course is for
Engineers deploying models on limited hardware.
What you will learn
- Quantise a model to 8, 4 and lower bits
- Distil a large model into a small one
- Prune a network without wrecking it
- Measure the quality you gave up
Syllabus
- Module 1: Quantisation methods
- Module 2: 1-bit models (BitNet)
- Module 3: Knowledge distillation
- Module 4: Pruning
- Module 5: Quality and speed measurement
Final project
Every module ends in hands-on practice, and the course ends with a project you build and present. Your certificate names that project.
Before you start
Deep learning.