This is still unfinished, I am currently working on it.
Unofficial code source for (Kaji and Kobayashi, 2017)
Pytorch implementation of incremental skip-gram model with negative sampling (Kaji and Kobayashi, 2017).
Please cite (Kaji and Kobayashi, 2017) when using this code to publish a paper.
@inproceedings{kaji17,
author = {Nobuhiro Kaji and Hayato Kobayashi},
title = {Incremental Skip-gram Model with Negative Sampling},
booktitle = {Proceedings of EMNLP},
year = {2017},
pages = {363--371}
}
arXiv version is also available here.