English

ByT5: Towards a token-free future with pre-trained byte-to-byte models

Computation and Language 2022-03-09 v3

Abstract

Most widely-used pre-trained language models operate on sequences of tokens corresponding to word or subword units. By comparison, token-free models that operate directly on raw text (bytes or characters) have many benefits: they can process text in any language out of the box, they are more robust to noise, and they minimize technical debt by removing complex and error-prone text preprocessing pipelines. Since byte or character sequences are longer than token sequences, past work on token-free models has often introduced new model architectures designed to amortize the cost of operating directly on raw text. In this paper, we show that a standard Transformer architecture can be used with minimal modifications to process byte sequences. We characterize the trade-offs in terms of parameter count, training FLOPs, and inference speed, and show that byte-level models are competitive with their token-level counterparts. We also demonstrate that byte-level models are significantly more robust to noise and perform better on tasks that are sensitive to spelling and pronunciation. As part of our contribution, we release a new set of pre-trained byte-level Transformer models based on the T5 architecture, as well as all code and data used in our experiments.

Keywords

Cite

@article{arxiv.2105.13626,
  title  = {ByT5: Towards a token-free future with pre-trained byte-to-byte models},
  author = {Linting Xue and Aditya Barua and Noah Constant and Rami Al-Rfou and Sharan Narang and Mihir Kale and Adam Roberts and Colin Raffel},
  journal= {arXiv preprint arXiv:2105.13626},
  year   = {2022}
}

Comments

To be published in TACL 2022

R2 v1 2026-06-24T02:33:31.538Z