Skip to content

wasimusu/MachineRC

Folders and files

NameName
Last commit message
Last commit date

Latest commit

b9a9d5e · May 1, 2019

History

71 Commits
Apr 18, 2019
Apr 26, 2019
Apr 16, 2019
Apr 17, 2019
Apr 17, 2019
Apr 13, 2019
Apr 18, 2019
May 1, 2019
Apr 17, 2019
Apr 30, 2019
Apr 18, 2019
Apr 19, 2019
Apr 18, 2019
Apr 17, 2019
Apr 17, 2019
Apr 12, 2019
Apr 16, 2019
Apr 16, 2019
Apr 16, 2019
Apr 30, 2019
Apr 17, 2019

Repository files navigation

PyTorch Implementation of Dynamic Coattention Network

Implementation of the paper Dynamic Coattention Network https://arxiv.org/pdf/1611.01604.pdf

Improvement ideas

  • Use layer normalization

Parts of the architecture explained in brief

Encoder

Encodes both question and context

Coattention Network

Combines attention of question with context

Highway Maxout Network

Determines possible start and end points

Dynamic Decoding

Determines start and end points

What do the py files do

  • config.py contains all the configuration

  • baseline.py contains a baseline architecture based on tfidf and cosine distance

  • vanillaQA.py contains baseline neural network architecture that might possibly work

  • squad.py contains data parser for Squad Dataset

  • setup.py - you need to run this after installing requirements to download data for nltk

  • networks package has all of the networks in separate class for testing purpose