Kitaev et al. (2019)#

Publication#

Constituency Parsing with a Self-Attentive Encoder

Multilingual Constituency Parsing with Self-Attention and Pre-Training

Repositories#

https://github.com/nikitakit/self-attentive-parser

Available Models#

  • Benepar

    • Description: A wrapper around the Benepar parser

    • Name: kitaev2019-benepar

    • Usage:

      from repro.models.kitaev2019 import Benepar
model = Benepar()
trees = model.predict("The time for action is now.")

      trees is a list of strings which contains the serialized parse trees for the input text.

Implementation Notes#

  • The input text does not have to be a single sentence. A parse tree will be returned for each one based on the library’s sentence splitting logic.

Docker Information#

  • Image name: danieldeutsch/kitaev2019

  • Build command:

    repro setup kitaev2019 [--models <model-name>+] [--silent]

    The --models argument specifies which pretrained parsing models should be included in the Docker image. By default, only the benepar_en3 model is included. The list of available models can be found here.

  • Requires network: No

Testing#

repro setup kitaev2019
pytest models/kitaev2019/tests

Status#

  • [x] Regression unit tests pass

  • [x] Correctness unit tests pass
    We verify the parser returns the example from the Github repo. See here.

  • [ ] Model runs on full test dataset
    Not tested

  • [ ] Predictions approximately replicate results reported in the paper
    Not tested

  • [ ] Predictions exactly replicate results reported in the paper
    Not tested

Changelog#