# Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context

## TL;DR

Transformer-XL introduces a novel neural architecture that extends the context length in language modeling, capturing longer dependencies and improving performance significantly over previous models while being computationally more efficient.

## Contribution

It proposes a segment-level recurrence and a new positional encoding scheme, enabling dependency learning beyond fixed-length contexts without losing temporal coherence.

## Key findings

- Learns dependencies 80% longer than RNNs and 450% longer than vanilla Transformers.
- Achieves state-of-the-art perplexity scores on multiple datasets.
- Up to 1,800+ times faster than vanilla Transformers during evaluation.

## Abstract

Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch.