# GRU

A recurrent unit with fewer gates than an LSTM, often similarly accurate and faster to train.

## What it is used for

Time-series and sequence modelling with tighter compute budgets.