Home Knowledge Base LoRA (Low-Rank Adaptation)

LoRA (Low-Rank Adaptation) is the parameter-efficient fine-tuning method that injects small trainable low-rank matrices into frozen pretrained model layers — enabling fine-tuning of billion-parameter models on consumer GPUs by training only 0.1-1% of total parameters while achieving 90-100% of full fine-tuning quality, democratizing LLM customization.

Core Idea

Why It Works

Parameter Efficiency

ModelFull FT ParamsLoRA (r=16)Reduction
LLaMA-7B6.7B~4M1675x
LLaMA-13B13B~6.5M2000x
LLaMA-70B70B~33M2121x

Memory Savings

QLoRA (Quantized LoRA)

Practical Configuration

ParameterTypical ValueNotes
Rank (r)8-64Higher = more capacity, more params
Alpha (α)16-32Scaling factor, often set to 2×rank
Target modulesq_proj, v_proj (attention)Can also target k_proj, o_proj, FFN
Dropout0.05-0.1On LoRA layers
Learning rate1e-4 to 3e-4Higher than full fine-tuning

LoRA Variants

Merging and Serving

LoRA is the technique that made LLM fine-tuning accessible to everyone — by reducing the hardware requirements from a cluster of A100s to a single consumer GPU, it enabled the explosion of open-source fine-tuned models and custom AI applications.

loralow rank adaptationqloraparameter efficient fine tuningpeft adapter

Explore 500+ Semiconductor & AI Topics

From EUV lithography to CUDA optimization — search the full knowledge base or chat with our AI assistant.