• / (root): Contain executable python files (create embedding, extract feature) and training/testing jupyter notebooks.
  • data/: Contains training and testing data.
  • helper_files/: Contains files (not code) that is needed by the system e.g. word embedding
  • model builders/: Builder class for singleton classifier and coreference classifier model
  • models/: Trained singleton classifier and coreference classifier models.
  • tensor_builders/: Helper file for building Keras tensors.
  • utils/: Helper codes e.g. data structures, coreference scorers, markable clusterers, training instance generator, etc.

This repository uses Git Large File System (LFS) to store (big) files inside data/ and models/ folder. To clone this repository, you should install Git LFS first. After that, you can clone this repository just like usual using git clone. Alternatively, you can use git lfs clone command to parallelize the downloads and reduce the download time. For further information, you can check Cloning an existing Git LFS repository.