Quantization Fundamentals with Hugging Face - Syllabus

• Introduction
  • Overview of quantization and its benefits for generative AI
  • Course objectives and key concepts
• Introduction to Quantization
  • Understanding the need for model compression
  • Basics of quantization techniques and applications
• Linear Quantization
  • Implementing linear quantization with the Quanto library
  • Compressing open source models using linear quantization
• Downcasting with Transformers
  • Applying downcasting to reduce model size
  • Using BFloat16 data type for optimized model performance
• Quantizing Multimodal Models
  • Applying quantization to multimodal models
  • Practicing quantization for language and vision models
• Quantization in Practice
  • Hands-on practice with open source models
  • Evaluating model performance after quantization
• Conclusion
  • Recap of quantization techniques and course highlights
  • Next steps for optimizing generative AI models