FastText

FastText is an open-source, free, and lightweight library developed by Facebook for efficient learning of word representations and text classification. It is designed to work on standard, generic hardware and supports model size reduction for deployment on mobile devices.

Features

• Text Representation Learning: Learns high-quality word embeddings (vector representations) from raw text data.
• Text Classification: Provides tools for fast and efficient text classification.
• Pre-trained Word Vectors: Offers English word vectors pre-trained on web crawl and Wikipedia data.
• Multi-lingual Support: Pre-trained models available for 157 different languages.
• Model Compression: Models can be reduced in size to fit on resource-constrained devices.
• Subword Information: Enriches word vectors with subword information for improved handling of rare and out-of-vocabulary words.
• Efficient on Generic Hardware: Designed to run efficiently on standard hardware without requiring GPUs.
• API and Tutorials: Provides a command-line API and comprehensive tutorials for ease of use.

Category

• SDKs & Libraries

Tags

open-source, vector-embeddings, semantic-search, machine-learning

Pricing

• Free and open-source