Course Overview
Learn to customize language models for your specific domain and use case. From parameter-efficient fine-tuning to building specialized small language models that run locally.
1
Fine-Tuning Fundamentals
Understanding when to fine-tune vs. use RAG. Data preparation, evaluation metrics, and avoiding catastrophic forgetting.
Data Prep
Evaluation
Benchmarks
2
PEFT Techniques
LoRA, QLoRA, Adapters, and Prefix Tuning. When to use each technique and how to configure them.
LoRA
QLoRA
Adapters
3
Small Language Models
Training and deploying SLMs for edge computing. Distillation, quantization, and optimization.
Distillation
Quantization
GGUF
4
MLOps for LLMs
Production deployment, monitoring, versioning, and continuous improvement of fine-tuned models.
MLflow
Weights & Biases
Model Registry
Hands-On Exercises
- Prepare a high-quality dataset for fine-tuning
- Fine-tune Llama with LoRA on domain-specific data
- Implement QLoRA for memory-efficient training
- Distill a large model into a smaller, faster version
- Quantize and deploy a model with llama.cpp
- Set up MLOps pipeline with experiment tracking and model versioning
Ready to Transform Your Team?
Contact us to discuss your training needs and schedule a consultation.
Get in Touch