Models, reasoning & training · Concept

QLoRA

Also called: quantized low-rank adaptation

Parameter-efficient adaptation using low-rank updates with a quantized base model.