{
 "title": "LoRA 低秩適應（Low-Rank Adaptation (LoRA)）是什麼？",
 "site": "AI 與科技名詞白話解釋",
 "canonical": "https://glossary.penguindriver.com/t/lora",
 "markdown": "https://glossary.penguindriver.com/t/lora.md",
 "summary": "LoRA（低秩適應）是一種參數高效的微調方法，凍結預訓練模型的權重，只在 Transformer 各層加入可訓練的小型低秩矩陣。微軟研究者 2021 年的論文指出，以 GPT-3 1,750 億參數為例，LoRA 可把需要訓練的參數減少 1 萬倍、GPU 記憶體需求降為三分之一，效果與完整微調相當或更好。",
 "date_modified": "2026-09-28",
 "retrieved": "2026-09-29",
 "content_sha256": "cc0c601da6a7032066d0208f7a6505884a686d739b1b6a8720be162539fa5429",
 "fingerprint_basis": "Markdown 版全文，移除贊助行與開頭 front matter，去除頭尾空白",
 "verify": {
  "text_snapshot": "https://glossary.penguindriver.com/cite?path=%2Ft%2Flora&format=text",
  "how": "下載 text_snapshot 後計算 SHA-256，應與 content_sha256 相同；內容更新時指紋會改變"
 },
 "sources": [
  {
   "name": "Hu et al., LoRA: Low-Rank Adaptation of Large Language Models（arXiv 2106.09685）",
   "url": "https://arxiv.org/abs/2106.09685"
  }
 ],
 "source_count": 1,
 "cite_as": {
  "zh": "LoRA 低秩適應（Low-Rank Adaptation (LoRA)）是什麼？｜AI 與科技名詞白話解釋。https://glossary.penguindriver.com/t/lora（資料日期 2026-09-28，擷取 2026-09-29）",
  "apa": "AI 與科技名詞白話解釋. (2026). LoRA 低秩適應（Low-Rank Adaptation (LoRA)）是什麼？. Retrieved 2026-09-29, from https://glossary.penguindriver.com/t/lora",
  "markdown": "[LoRA 低秩適應（Low-Rank Adaptation (LoRA)）是什麼？](https://glossary.penguindriver.com/t/lora)（AI 與科技名詞白話解釋，2026-09-28）"
 },
 "license": "可引用，請附上正式網址與資料日期"
}