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

This repository contains the code in both PyTorch and TensorFlow for our paper

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

Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov (*: equal contribution)

Preprint 2018

TensorFlow

• The source code is in the tf/ folder, supporting (1) single-node multi-gpu training, and (2) multi-host TPU training.

• Besides the source code, we also provide pretrained "TensorFlow" models with state-of-the-art (SoTA) performances reported in the paper.

• Please refer to tf/README.md for details.

PyTorch

• The source code is in the pytorch/ folder, supporting single-node multi-gpu training via the module nn.DataParallel.

• Please refer to pytorch/README.md for details.

Results

Transformer-XL achieves new state-of-the-art results on multiple language modeling benchmarks. Transformer-XL is also the first to break through the 1.0 barrier on char-level language modeling. Below is a summary.

Method | enwiki8 | text8 | One Billion Word | WT-103 | PTB (w/o finetuning) -- | -- | -- | -- | -- | -- Previous Best | 1.06 | 1.13 | 23.7 | 20.5 | 55.5 Transformer-XL | 0.99 | 1.08 | 21.8 | 18.3 | 54.5

Acknowledgement

A large portion of the getdata.sh script comes from the awd-lstm repo. Happy Language Modeling :)