# Int4 QLoRA NF4 Internals: Double Quantization + Paged Optimizer + Bitsandbytes Source Tour

> Source: https://sukruyusufkaya.com/en/learn/fine-tuning-cookbook/ftc-qlora-nf4-double-quantization-internals
> Updated: 2026-05-14T14:42:57.463Z
> Category: Fine-Tuning Cookbook (Model-by-Model)
> Module: Part X — Quantization Engineering
**TLDR:** NF4 (4-bit NormalFloat) — the core of QLoRA. Optimal 4-bit quantization for normally-distributed weights. Double-quantization (also quantize the scale tensor) for additional 0.4 bit/param savings. Paged AdamW (overflow to CPU RAM). Bitsandbytes source-code tour.