* Support of encoder module for character-based models; * Implementation of document-term matrix (DTM).
0.5.9
---------------------------------
* Support of Poincare embedding; * Code optimization; * Script `ShortTextWord2VecSimilarity` updated to `ShortTextWordEmbedSimilarity`.
0.5.8
--------------------------------
* Removed most explicit user-specification of `vecsize` for given word-embedding models; * Removed old namespace for topic models (no more backward compatibility). * Integration of [FastText](https://github.com/facebookresearch/fastText).
0.5.7
--------------------------------
* Removed most explicit user-specification of `vecsize` for given word-embedding models; * Removed old namespace for topic models (hence no more backward compatibility).
0.5.6
--------------------------------
* Updated the neural network framework due to the change in `gensim` API.