GitHub - vec2text/vec2text: utilities for decoding deep representations (like sentence embeddings) back to text
utilities for decoding deep representations (like sentence embeddings) back to text - vec2text/vec2text
Pretrained Word Embeddings Resource
Summary
The page lists several pre‑trained English word vector files released with fastText, including models trained on Wikipedia, Common Crawl, and news data. It describes file formats, provides a Python loading example, and includes licensing and citation information.
Highlights
- Multiple fastText English vectors (wiki‑news, crawl) with and without subword information.
- Files are in the standard fastText text format; first line gives vocab size and vector dimension.
- Download links include 300‑dimensional 1M‑word and 2M‑word models covering billions of tokens.
- Simple Python code snippet shows how to load vectors efficiently.
- Distributed under CC BY‑SA 3.0 with a recommended citation.
auto-generated
Context
Audience
Researchers, data scientists, and engineers working on NLP tasks who need high‑quality English word embeddings
DomainMachine Learning / Natural Language Processing
Formatdownloadable .vec zip archives and plain‑text documentation
Accessfree online download without registration
Source Overview
The page provides several pre-trained English word vector files generated with fastText, including models trained on Wikipedia, Common Crawl, and news datasets.
Files are available with and without subword information, supporting robust handling of both common and rare words. All vectors follow the fastText text format and are distributed under the Creative Commons Attribution-Share-Alike 3.0 license.
Topics
Related
fastText libraryGloVe word vectorsWord2Vec embeddingsNLTK / spaCy integration tutorialsLREC 2018 citation for fastText pre‑training
