> ## Documentation Index
> Fetch the complete documentation index at: https://nixtlaverse.nixtla.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Install | MLForecast

> Instructions to install the package from different sources.

## Released versions

### PyPI

#### Latest release

To install the latest release of mlforecast from
[PyPI](https://pypi.org/project/mlforecast/) you just have to run the
following in a terminal:

`pip install mlforecast`

#### Specific version

If you want a specific version you can include a filter, for example:

* `pip install "mlforecast==0.3.0"` to install the 0.3.0 version
* `pip install "mlforecast<0.4.0"` to install any version prior to
  0.4.0

#### Extras

**polars**

Using polars dataframes: `pip install "mlforecast[polars]"`

**Saving to remote storages**

If you want to save your forecast artifacts to a remote storage like S3
or GCS you can use the following extras:

* Saving to S3: `pip install "mlforecast[aws]"`
* Saving to Google Cloud Storage: `pip install "mlforecast[gcp]"`
* Saving to Azure Data Lake: `pip install "mlforecast[azure]"`

**Distributed training**

If you want to perform distributed training you can use either dask, ray
or spark. Once you know which framework you want to use you can include
its extra:

* dask: `pip install "mlforecast[dask]"`
* ray: `pip install "mlforecast[ray]"`
* spark: `pip install "mlforecast[spark]"`

### Conda

#### Latest release

The mlforecast package is also published to
[conda-forge](https://anaconda.org/conda-forge/mlforecast), which you
can install by running the following in a terminal:

`conda install -c conda-forge mlforecast`

Note that this happens about a day later after it is published to PyPI,
so you may have to wait to get the latest release.

#### Specific version

If you want a specific version you can include a filter, for example:

* `conda install -c conda-forge "mlforecast==0.3.0"` to install the
  0.3.0 version
* `conda install -c conda-forge "mlforecast<0.4.0"` to install any
  version prior to 0.4.0

## Development version

If you want to try out a new feature that hasn’t made it into a release
yet you have the following options:

* Install from github:
  `pip install git+https://github.com/Nixtla/mlforecast`
* Clone and install:
  `git clone https://github.com/Nixtla/mlforecast mlforecast-dev && pip install mlforecast-dev/`,
  which will install the version from the current main branch.