Metadata-Version: 2.1
Name: fasttext_reducer
Version: 0.1.3
Summary: Lightweight package that allows for downloading fasttext models to any location and reducing their dimensions through a one-liner.
Author-email: Ramtin Yazdanian <ramtin.yazdanian@epfl.ch>
Project-URL: Homepage, https://github.com/RamtinYazdanian/fasttext-reducer
Project-URL: Bug Tracker, https://github.com/RamtinYazdanian/fasttext-reducer/issues
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: fasttext-wheel

# fasttext-reducer
A tiny package (and standalone script) for downloading any pretrained fasttext word vector model to any desired location and then reducing its number of dimensions.

## Installation
Either run the following to install from PyPI:
```
pip install fasttext-reducer
```

Or clone the repository and install the package using `pip install .`

## Example usage
To download the English pretrained model, reduce it to 30 dimensions, and save it to `/Users/admin/`, you can run the following command:
```
fasttext-reduce --root_dir /Users/admin --lang en --dim 30
```
You can also use the package within Python:
```
from fasttext_reducer.reduce_fasttext_models import reduce_fasttext_models

reduce_fasttext_models('/Users/admin', 'en', 30)
```
