Skip to content

Anirudh-C/neural-constituent-parser

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Neural Constituent Parser

Constituency parsing aims to extract a constituency-based parse tree from a sentence that represents its syntactic structure according to a phrase structure grammar.

We use a 2 element span (i,j,k,l) (representing 2 child phrases in the phrase tree) and an encoding of the rule to score the spans on the sentence. We maximize the log-likelihood of the probability of a phrase tree given a sentence which is proportional to the exponential of the sum of the scores for the spans in the tree.

We also bound the score function using a RELu6 unit.

Usage

First make sure all the dependencies are installed

$ pip3 install -r requirements.txt

To train a model run train.py.

Run the following to see the arguments that can be passed to train.py

$ python3 train.py -h

About

A Neural Constituent parser for English

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages