LoRA
Full Form: Low-Rank Adaptation
Category: AI Techniques
📖 Definition
LoRA is an efficient technique for fine-tuning large models by adding small trainable matrices instead of retraining the entire model. It's much faster and requires less compute.
🔑 Key Points
- Fine-tunes models without updating all parameters
- Much faster and cheaper than full fine-tuning
- Popular for customizing image and language models
- Enables multiple specialized versions of one model
💡 Why It Matters
LoRA democratized AI customization. It lets hobbyists and businesses customize powerful models on limited hardware.