{
  "id": "2404.08080",
  "version": "v1",
  "published": "2024-04-11T18:35:49.000Z",
  "updated": "2024-04-11T18:35:49.000Z",
  "title": "Variance-reduced Zeroth-Order Methods for Fine-Tuning Language Models",
  "authors": [
    "Tanmay Gautam",
    "Youngsuk Park",
    "Hao Zhou",
    "Parameswaran Raman",
    "Wooseok Ha"
  ],
  "comment": "29 pages, 25 tables, 9 figures",
  "categories": [
    "cs.LG",
    "cs.AI",
    "cs.CL",
    "math.OC"
  ],
  "abstract": "Fine-tuning language models (LMs) has demonstrated success in a wide array of downstream tasks. However, as LMs are scaled up, the memory requirements for backpropagation become prohibitively high. Zeroth-order (ZO) optimization methods can leverage memory-efficient forward passes to estimate gradients. More recently, MeZO, an adaptation of ZO-SGD, has been shown to consistently outperform zero-shot and in-context learning when combined with suitable task prompts. In this work, we couple ZO methods with variance reduction techniques to enhance stability and convergence for inference-based LM fine-tuning. We introduce Memory-Efficient Zeroth-Order Stochastic Variance-Reduced Gradient (MeZO-SVRG) and demonstrate its efficacy across multiple LM fine-tuning tasks, eliminating the reliance on task-specific prompts. Evaluated across a range of both masked and autoregressive LMs on benchmark GLUE tasks, MeZO-SVRG outperforms MeZO with up to 20% increase in test accuracies in both full- and partial-parameter fine-tuning settings. MeZO-SVRG benefits from reduced computation time as it often surpasses MeZO's peak test accuracy with a $2\\times$ reduction in GPU-hours. MeZO-SVRG significantly reduces the required memory footprint compared to first-order SGD, i.e. by $2\\times$ for autoregressive models. Our experiments highlight that MeZO-SVRG's memory savings progressively improve compared to SGD with larger batch sizes.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2024-04-11T18:35:49.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "fine-tuning language models",
      "variance-reduced zeroth-order methods",
      "zeroth-order stochastic variance-reduced gradient",
      "surpasses mezos peak test accuracy"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 29,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}