Andras7 / word2vec-pytorch

Extremely simple and fast word2vec implementation with Negative Sampling + Sub-sampling

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Word2vec Pytorch

Fast word2vec implementation at competitive speed compared with fasttext. The slowest part is the python data loader. Indeed, Python isn't the fastest programming language, maybe you can improve the code :)

Advantages

  • Easy to understand, solid code
  • Easy to extend for new experiments
  • You can try advanced learning optimizers, with new learning technics
  • GPU support

Supported features

  • Skip-gram
  • Batch update
  • Cosine Annealing
  • Negative Sampling
  • Sub-sampling of frequent word

About

Extremely simple and fast word2vec implementation with Negative Sampling + Sub-sampling


Languages

Language:Python 100.0%