MDSection 24 of 38 · 7 subtopics
LoRA
Low-rank adaptation: freeze weights, train tiny matrices. 100x fewer trainable parameters.
Learning path
Work through each subtopic in order. Click any file below to open its lesson. Track your progress with the checkbox at the bottom.
Subtopics
Each subtopic includes a concise explanation, code examples, and references.
01
Low-Rank Adaptation
Frozen model + low-rank adapter diagram.
low-rank-adaptation
02
Frozen Weights
Model parameter map.
frozen-weights
03
Trainable Adapters
Adapter overlay.
trainable-adapters
04
Low-Rank Matrices
A × B matrix reconstruction visualizer.
low-rank-matrices
05
LoRA Rank
Rank slider.
lora-rank
06
LoRA Merging
Base + adapter merge animation.
lora-merging
07
QLoRA
Quantized base model + LoRA adapter.
qlora