Log In Sign Up

FNet: Mixing Tokens with Fourier Transforms

by   James Lee-Thorp, et al.

We show that Transformer encoder architectures can be massively sped up, with limited accuracy costs, by replacing the self-attention sublayers with simple linear transformations that "mix" input tokens. These linear transformations, along with simple nonlinearities in feed-forward layers, are sufficient to model semantic relationships in several text classification tasks. Perhaps most surprisingly, we find that replacing the self-attention sublayer in a Transformer encoder with a standard, unparameterized Fourier Transform achieves 92 to seven times faster on GPUs and twice as fast on TPUs. The resulting model, which we name FNet, scales very efficiently to long inputs, matching the accuracy of the most accurate "efficient" Transformers on the Long Range Arena benchmark, but training and running faster across all sequence lengths on GPUs and relatively shorter sequence lengths on TPUs. Finally, FNet has a light memory footprint and is particularly efficient at smaller model sizes: for a fixed speed and accuracy budget, small FNet models outperform Transformer counterparts.


page 1

page 2

page 3

page 4


FNetAR: Mixing Tokens with Autoregressive Fourier Transforms

In this note we examine the autoregressive generalization of the FNet al...

FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks

Transformers achieve remarkable performance in various domains, includin...

PoNet: Pooling Network for Efficient Token Mixing in Long Sequences

Transformer-based models have achieved great success in various NLP, vis...

When Can Self-Attention Be Replaced by Feed Forward Layers?

Recently, self-attention models such as Transformers have given competit...

Waveformer: Linear-Time Attention with Forward and Backward Wavelet Transform

We propose Waveformer that learns attention mechanism in the wavelet coe...

Block-Recurrent Transformers

We introduce the Block-Recurrent Transformer, which applies a transforme...

ProFormer: Towards On-Device LSH Projection Based Transformers

At the heart of text based neural models lay word representations, which...