Skip to content
NerdLevelTech
Intermediate

Fine-tuning LLMs: LoRA, QLoRA & PEFT

Master parameter-efficient fine-tuning techniques to customize large language models for your domain. Learn LoRA, QLoRA, DPO alignment, and deploy fine-tuned models to production with Unsloth and Ollama.

72 min24 lessons6 modulesAI & ML

What you'll learn

  • Apply LoRA and QLoRA techniques for memory-efficient fine-tuning
  • Prepare and format datasets for domain-specific LLM training
  • Implement DPO alignment for improved model responses
  • Deploy fine-tuned models locally using Unsloth and Ollama
  • Evaluate fine-tuned model performance against baseline metrics

Course Content

  1. Understanding Fine-tuning

    4 lessons · 12 min
    • Why Fine-tune LLMs?No signup3 min
    • Locked until you enrollTypes of Fine-tuning3 min
    • Locked until you enrollPEFT Methods Explained3 min
    • Locked until you enrollChoosing Your Approach3 min
    • Quiz: Module 1: Understanding Fine-tuning

  2. Dataset Preparation

    4 lessons · 12 min
    • Locked until you enrollInstruction Tuning Datasets3 min
    • Locked until you enrollCreating Training Data3 min
    • Locked until you enrollData Quality and Cleaning3 min
    • Locked until you enrollPreference Data for DPO3 min
    • Quiz: Module 2: Dataset Preparation

  3. LoRA & QLoRA in Practice

    4 lessons · 12 min
    • Locked until you enrollSetting Up the Environment3 min
    • Locked until you enrollLoRA Configuration3 min
    • Locked until you enrollQLoRA: 4-bit Fine-tuning3 min
    • Locked until you enrollYour First Fine-tune3 min
    • Quiz: Module 3: LoRA & QLoRA in Practice

  4. Training with Unsloth

    4 lessons · 12 min
    • Locked until you enrollWhy Unsloth?3 min
    • Locked until you enrollUnsloth Setup3 min
    • Locked until you enrollTraining with Unsloth3 min
    • Locked until you enrollExporting Your Model3 min
    • Quiz: Module 4: Training with Unsloth

  5. Alignment with DPO

    4 lessons · 12 min
    • Locked until you enrollDPO vs RLHF3 min
    • Locked until you enrollPreparing DPO Data3 min
    • Locked until you enrollDPOTrainer in Practice3 min
    • Locked until you enrollCombining SFT + DPO3 min
    • Quiz: Module 5: Alignment with DPO

  6. Evaluation & Deployment

    4 lessons · 12 min
    • Locked until you enrollEvaluating Fine-tuned Models3 min
    • Locked until you enrollCommon Issues & Fixes3 min
    • Locked until you enrollDeploying to Ollama3 min
    • Locked until you enrollNext Steps3 min
    • Quiz: Module 6: Evaluation & Deployment