Leader: Prefixing a Length for Faster Word Vector Serialization
This work addresses a practical problem for researchers and practitioners using pre-trained word embeddings by improving serialization speed and file size, though it is incremental as it builds on existing binary formats.
The authors tackled the inefficiency of existing word embedding file formats by introducing the Leader format, which uses a word length prefix to achieve faster reads while maintaining small file sizes, and they provided a library and tools for format conversion.
Two competing file formats have become the de facto standards for distributing pre-trained word embeddings. Both are named after the most popular pre-trained embeddings that are distributed in that format. The GloVe format is an entirely text based format that suffers from huge file sizes and slow reads, and the word2vec format is a smaller binary format that mixes a textual representation of words with a binary representation of the vectors themselves. Both formats have problems that we solve with a new format we call the Leader format. We include a word length prefix for faster reads while maintaining the smaller file size a binary format offers. We also created a minimalist library to facilitate the reading and writing of various word vector formats, as well as tools for converting pre-trained embeddings to our new Leader format.