Installation#
GluonTS is available from PyPi via:
pip install gluonts
Caution
GluonTS uses Semantic Versioning for managing versions.
Since the library is actively developed we use v0 as the major version.
We plan to release a new minor version at the end of each quarter. The current planned releases can be found on GitHub.
Version Guarantees
Breaking changes are only introduced with a new minor release. Bug fixes and minor improvements are provided for the current release and are published on demand.
For production usage, we suggest restricting the version when installing GluonTS:
pip install gluonts==0.12.*
Optional and Extra Dependencies#
Important
GluonTS uses a minimal dependency model.
This means that to use most models and features additional dependencies need to be installed.
Python has the notion of extras – dependencies that can be optionally installed to unlock certain features of a package.
When installing a package, they are passed via [...] after the package
name, but before the version specifier:
pip install "some-package[extra-1,extra-2]==version"
We make extensive use of optional dependencies in GluonTS to keep the amount of required dependencies minimal. To still allow users to opt-in to certain features, we expose many extra dependencies.
For example, we offer support for reading and writing Arrow and Parquet based
datasets using Apache Arrow. However, it is a
hefty dependency to require, especially if one has no need for it. Thus, we
offer the arrow-extra, which installs the required packages and can be
simply enabled using:
pip install "gluonts[arrow]"
Models#
PyTorch#
Models written using PyTorch are available via the
gluonts.torch subpackage.
In addition to PyTorch we require PyTorch Lightning to be installed as well.
Both required dependencies are included in the torch-extra:
pip install "gluonts[torch]"
MXNet#
MXNet based models require a version of mxnet to be installed.
Note
MXNet provides different package for CPU and GPU usages. Please refer to its documentation to select the right version fitting your use-case.
The mxnet-extra will install a CPU-only version:
pip install "gluonts[mxnet]"
3rd Party#
R-Forecast#
GluonTS includes a thin wrapper for calling the R forecast package.
In order to use it you need to install R and
install the forecast package:
R -e 'install.packages(c("forecast", "nnfor"), repos="https://cloud.r-project.org")'
You will also need to install R-extra dependencies with:
pip install "gluonts[R]"
Datasets#
JSON#
Since Python’s build in json package is known to be relatively slow, we use
faster implementations if available: orjson (recommended) and ujson.
You can install orjson via:
pip install orjson
Hint
GluonTS will emit a warning if neither orjson nor ujson are installed.
There is no functional difference between the different implementations, but
especially when working with larger datasets, performance can be notably
impacted when relying on the default json package.
Arrow#
GluonTS support Parquet files
using PyArrow.
Further, arrow’s custom data formats are also supported.
To use these, either install the pyarrow package or use the
arrow-extra:
pip install "gluonts[arrow]"