English

Random-LTD: Random and Layerwise Token Dropping Brings Efficient Training for Large-scale Transformers

Computation and Language 2022-11-22 v1 Machine Learning

Abstract

Large-scale transformer models have become the de-facto architectures for various machine learning applications, e.g., CV and NLP. However, those large models also introduce prohibitive training costs. To mitigate this issue, we propose a novel random and layerwise token dropping method (random-LTD), which skips the computation of a subset of the input tokens at all middle layers. Particularly, random-LTD achieves considerable speedups and comparable accuracy as the standard training baseline. Compared to other token dropping methods, random-LTD does not require (1) any importance score-based metrics, (2) any special token treatment (e.g., [CLS]), and (3) many layers in full sequence length training except the first and the last layers. Besides, a new LayerToken learning rate schedule is proposed for pretraining problems that resolve the heavy tuning requirement for our proposed training mechanism. Finally, we demonstrate that random-LTD can be applied to broader applications, including GPT and BERT pretraining as well as ViT and GPT finetuning tasks. Our results show that random-LTD can save about 33.3% theoretical compute cost and 25.6% wall-clock training time while achieving similar zero-shot evaluations on GPT-31.3B as compared to baseline.

Keywords

Cite

@article{arxiv.2211.11586,
  title  = {Random-LTD: Random and Layerwise Token Dropping Brings Efficient Training for Large-scale Transformers},
  author = {Zhewei Yao and Xiaoxia Wu and Conglong Li and Connor Holmes and Minjia Zhang and Cheng Li and Yuxiong He},
  journal= {arXiv preprint arXiv:2211.11586},
  year   = {2022}
}

Comments

22 pages

R2 v1 2026-06-28T06:23:10.186Z