snakefusion is a Python package for reading, writing, and using finalfusion, fastText, floret, GloVe, and word2vec embeddings. This package is a thin wrapper around the Rust finalfusion crate.

snakefusion supports the same types of embeddings as finalfusion:

• Vocabulary:
  • No subwords
  • Subwords
• Embedding matrix:
  • Array
  • Memory-mapped
  • Quantized
• Format:
  • fastText
  • finalfusion
  • floret
  • GloVe
  • word2vec

Building snakefusion from source requires a Rust toolchain that is installed through rustup and setuptools-rust:

$ pip install --upgrade setuptools-rust

You can then build and install snakefusion in your environment:

$ pip install .

Read the snakefusion documentation for a quickstart and API reference.

You use finalfrontier to train new embeddings, or download some pretrained embeddings.