Q

QLoRA

• Category: Math & training
• What people say: LoRA with a quantized base model.
• What it actually means: A parameter-efficient fine-tuning method that keeps a pretrained base model frozen in a low-bit quantized representation while training LoRA adapters with higher-precision computation where needed.
• Why it matters: It can reduce the memory needed to adapt large models, but savings and quality depend on model, rank, optimizer, sequence length, hardware, and implementation.
• Common confusion: QLoRA does not guarantee a particular memory footprint or a fixed quality gap from full fine-tuning.
• Learn it: Fine-Tuning and LoRA
• Sources: QLoRA paper
• Related terms: LoRA (Low-Rank Adaptation), Quantization, Fine-tuning

Quantization

• Category: Models & inference
• What people say: Storing or computing model values with fewer bits.
• What it actually means: Representing weights, activations, or caches with lower-precision formats to reduce memory, bandwidth, or compute cost. Methods differ in calibration, granularity, data type, and whether conversion happens before, during, or after training.
• Common confusion: Moving from one nominal bit width to another does not guarantee the same end-to-end memory or speed ratio because metadata, kernels, caches, and hardware support also matter.
• Related terms: QLoRA, Mixed Precision, Parameter