FitzGerald et al. (2018)#

Publication#

Large-Scale QA-SRL Parsing

Repositories#

https://github.com/nafitzgerald/nrl-qasrl

Available Models#

  • QA-SRL Parser

    • Description: A QA-SRL parser

    • Name: fitzgerald2018-qasrl-parser

    • Usage:

      from repro.models.fitzgerald2018 import QASRLParser
model = QASRLParser()
output = model.predict("The sentence to parse.")

      The output is a dictionary with the data output by the parser. See the original repository or our unit tests for an example.

Implementation Notes#

Docker Information#

  • Image name: danieldeutsch/fitzgerald2018

  • Build command:

    repro setup fitzgerald2018

  • Requires network: Yes, AllenNLP still sends a request even when a dataset is already cached locally.

Testing#

repro setup fitzgerald2018
pytest models/fitzgerald2018/tests

Status#

  • [x] Regression unit tests pass
    See here

  • [x] Correctness unit tests pass
    See here. The output is slightly different than what is expected in the Readme from the original repo, but it looks close enough.

  • [ ] 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#

v1.1#

  • The original script to download the model stopped working. The Dockerfile has been updated to directly download the model.