the.ai

Optimization / Regimes

verified

QLoRA

LoRA makes the trainable parameters small, but the frozen base model still has to sit in memory at full precision, and that is what actually decides whether you can fine-tune on the hardware you have. QLoRA quantizes the frozen base to four bits and trains the adapters on top of it, which is what puts fine-tuning a large model on a single accelerator.

Viz primitive · budget-splitbase-bits = 4

base-bits holds 80% of the budget; rest holds the remaining 20%.

Memory held by the frozen base against the trainable adapters beside it, in equal units. Drag the base precision and watch the base stay the majority at every setting — which is why quantizing it, not shrinking the adapter, is what moves the total.

4

Reviewed by opendroid · 2026-08-04