Skip to main content

Model Fine-tuning

LoRA, QLoRA, PEFT - customize models for your use cases.

Course Duration: 8 hours​

What You'll Learn​

  • Fine-tuning strategies
  • LoRA and QLoRA
  • PEFT methods
  • Dataset preparation
  • Training and evaluation
  • Deployment of fine-tuned models

Prerequisites​

  • LLM understanding
  • Python and PyTorch

Course Modules​

  1. Introduction to Fine-tuning
  2. Full Fine-tuning vs PEFT
  3. LoRA (Low-Rank Adaptation)
  4. QLoRA (Quantized LoRA)
  5. Dataset Preparation
  6. Training Configuration
  7. Hyperparameter Tuning
  8. Evaluation Metrics
  9. Model Merging
  10. GGUF and Quantization
  11. Deploying Fine-tuned Models
  12. Project: Custom Chatbot

Tools Covered​

  • HuggingFace Transformers
  • PEFT library
  • bitsandbytes
  • Axolotl
  • Unsloth