Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

Parameter-efficient fine-tuning

Parameter-efficient fine-tuning (PEFT) adapts a pretrained model by training only a small number of new or selected parameters, leaving the rest frozen. It cuts the memory, compute and storage that full fine-tuning needs.

Also called: PEFT

By Divya Sharma · Checked Nov 15, 2026 · How we review

Full fine-tuning updates every weight in a model. For a model with billions of parameters that needs several large GPUs, and each fine-tuned copy is as big as the original. PEFT methods avoid this by freezing the base model and training a small add-on.

How it works

The best-known method is LoRA, which adds small low-rank matrices next to existing weight matrices and trains only those. Other families add short trainable "soft prompts" to the input (prompt tuning, prefix tuning), insert small adapter layers between existing ones, or train only a subset of existing weights. QLoRA combines LoRA with a quantized base model so that training fits on a single consumer GPU for many model sizes.

The result is usually a file of a few megabytes to a few hundred megabytes that you load on top of the base model. One base model can serve many adapters, one per customer or task.

Why it matters

PEFT is why fine-tuning open models is practical on modest hardware. Libraries such as Hugging Face PEFT, Unsloth and LLaMA-Factory wrap these methods. Quality on a narrow task is often close to full fine-tuning.

Limits

PEFT changes behaviour and style well but is a weak way to add large amounts of new knowledge; retrieval-augmented generation is usually better for that. An adapter is tied to the exact base model it was trained on and will not work on a different one.

Related terms

LoRA
LoRA (low-rank adaptation) is a fine-tuning method that freezes a model's original weights and trains small low-rank matrices added to its layers, so far fewer parameters change and much less GPU memory is needed.
Fine-tuning
Fine-tuning is further training of a pretrained model on a smaller, task-specific dataset, so the model adopts a particular skill, format, tone or domain without being trained from scratch.
Supervised fine-tuning
Supervised fine-tuning (SFT) trains a pretrained model further on example prompts paired with the responses you want. It is the usual first step that turns a base model into one that follows instructions.
Quantization
Quantization is a technique that stores a model's weights, and sometimes activations, with fewer bits, such as 8 or 4 instead of 16, so the model needs less memory and often runs faster, at some cost in accuracy.
Model weights
Model weights are the learned numbers inside a neural network, set during training, that decide how it turns input into output. A checkpoint is a saved file or set of files containing those weights.

Sources