https://github.com/mlverse/tft

R implementation of Temporal Fusion Transformers
https://github.com/mlverse/tft

Last synced: 12 months ago
JSON representation

R implementation of Temporal Fusion Transformers

• Host: GitHub
• URL: https://github.com/mlverse/tft
• Owner: mlverse
• License: other
• Created: 2021-01-09T14:36:32.000Z (over 5 years ago)
• Default Branch: main
• Last Pushed: 2024-11-16T22:32:23.000Z (over 1 year ago)
• Last Synced: 2025-07-07T13:40:40.768Z (12 months ago)
• Language: R
• Homepage: https://mlverse.github.io/tft/
• Size: 4.39 MB
• Stars: 28
• Watchers: 5
• Forks: 9
• Open Issues: 14
• Metadata Files:
  • Readme: README.Rmd
  • License: LICENSE

Awesome Lists containing this project

README

          
---

output: github_document

---

```{r, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>",

  fig.path = "man/figures/README-",

  out.width = "100%"

)

```

# tft

[![R build status](https://github.com/mlverse/tft/workflows/R-CMD-check/badge.svg)](https://github.com/mlverse/tft/actions)

[![Lifecycle: experimental](https://img.shields.io/badge/lifecycle-experimental-orange.svg)](https://www.tidyverse.org/lifecycle/#experimental)

[![CRAN status](https://www.r-pkg.org/badges/version/tft)](https://CRAN.R-project.org/package=tft)

[![](https://cranlogs.r-pkg.org/badges/tft)](https://cran.r-project.org/package=tft)

[![Codecov test coverage](https://codecov.io/gh/mlverse/tft/branch/master/graph/badge.svg)](https://codecov.io/gh/mlverse/tft?branch=master)

> An R implementation of [tft: Temporal Fusion Transformer](https://arxiv.org/pdf/1912.09363.pdf).

The Temporal Fusion Transformer is a neural network architecture proposed by Bryan Lim et al. with the goal of making multi-horizon time series forecasts for

multiple time series in a single model.

The main difference between TFT and conventional forecasting methodologies is the

way its architecture allows encoding different types of input data that can exist

in forecasting problems. For instance, the model allows handling static covariates

and time varying (known and unknown) differently. tft also showed [promising benchmarks](https://ai.googleblog.com/2021/12/interpretable-deep-learning-for-time.html).

The code in this repository is heavily inspired in code from [akeskiner/Temporal_Fusion_Transform](https://github.com/akeskiner/Temporal_Fusion_Transform), [jdb78/pytorch-forecasting](https://github.com/jdb78/pytorch-forecasting) and

the original implementation [here](https://github.com/google-research/google-research/tree/master/tft).

## Installation

You can install the development version [GitHub](https://github.com/) with:

```r

# install.packages("remotes")

remotes::install_github("mlverse/tft")

```

Read the [Getting Started](https://mlverse.github.io/tft/articles/Getting-started.html) guide to fit your first

model with tft.