Word2Vec: Optimal Hyper-Parameters and Their Impact on NLP Downstream Tasks

Codes:

One source code file generates word2vec models while the other two are for downstream tasks (SA & NER).

W2V Model window8_CBoW_negativesampling:

download

License:

All codes and models are released under CCBY 4.0.

Contact:

Feel free to contact me for further enquiry.

How do I cite?

For now, cite the Arxiv paper:

@article{adewumi2020word2vec,
  title={Word2Vec: Optimal Hyper-Parameters and Their Impact on NLP Downstream Tasks},
  author={Adewumi, Tosin P and Liwicki, Foteini and Liwicki, Marcus},
  journal={arXiv preprint arXiv:2003.11645},
  year={2020}
}