Companion videos

Fine-tuning in 2026: when LoRA beats RAG, and how to do it without a cluster — companion videos

The article argues that most fine-tuning projects are still solving the wrong problem, but that LoRA on the right narrow task is one of the highest-leverage moves an engineer can make in 2026 — and that you no longer need a cluster to do it. Pair it with one long hands-on tutorial and one slower conceptual overview from a respected ML educator, and you should be able to decide both whether to fine-tune and how.

Primary pick

2:37:05

Fine Tuning LLM Models – Generative AI Course

freeCodeCamp.org

Long, theory-then-code course covering quantisation, LoRA, QLoRA, and full PEFT on Llama 2 and Gemma — on hardware most developers actually have. It is the closest thing to a "shadow somebody who has done this" experience on YouTube and lines up with the article's "you don't need a cluster" claim with concrete VRAM budgets.

What you should get from this: You will be able to run quantisation, LoRA and QLoRA fine-tunes end to end on hardware most developers already have.

Watch or know first: Python plus basic ML tooling comfort; the course targets hardware most developers already have.

AI Expert note: Model names, pricing and capabilities change quickly. The course's base models (Llama 2, Gemma) are dated, but the LoRA, QLoRA and quantisation techniques it teaches transfer directly to current open-weight models. Use this for the decision pattern, then verify current model behavior before adopting it.

Open video page

Also worth watching

58:45

Developing an LLM: Building, Training, Finetuning

Sebastian Raschka

Sebastian Raschka's slower walkthrough of where fine-tuning sits in the broader LLM training pipeline — instruction tuning, classification fine-tuning, parameter-efficient methods, and the trade-offs the article calls out before recommending LoRA. Good calibration before you start, especially if your team is debating whether fine-tuning is even the right step.

What you should get from this: Place fine-tuning inside the broader training pipeline and decide when it is better than prompting or RAG.

Watch or know first: General ML literacy; it's a conceptual walkthrough, so no code-along setup is needed.

AI Expert note: Model names, pricing and capabilities change quickly. Use this for the decision pattern, then verify current model behavior before adopting it.

Open video page