The models submodule contains implementations of various algorithms that can be used in addition to external packages to evaluate and develop new natural language processing systems. A description of which algorithms are used in each scenario can be found on this table
The following table summarizes each submodule.
| Submodule | Description |
|---|---|
| bert | This submodule includes the BERT-based models for sequence classification, token classification, and sequence encoding. |
| gensen | This submodule includes a distributed Pytorch implementation based on Horovod of learning general purpose distributed sentence representations via large scale multi-task learning by refactoring https://github.com/Maluuba/gensen |
| pretrained embeddings | This submodule provides utilities to download and extract pretrained word embeddings trained with Word2Vec, GloVe, fastText methods. |
| pytorch_modules | This submodule provides Pytorch modules like Gated Recurrent Unit with peepholes. |
| xlnet | This submodule includes the XLNet-based model for sequence classification. |