Pytorch implementations of various Deep NLP models in cs-224n(Stanford Univ)
Switch branches/tags
Nothing to show
Clone or download
Latest commit 7fec64d Jan 26, 2018
Permalink
Failed to load latest commit information.
images update readme Nov 9, 2017
notebooks fix some error Jan 26, 2018
script test Nov 5, 2017
.gitignore mkdir notebooks Oct 29, 2017
Corpus.pyx nmt Oct 8, 2017
LICENSE Create LICENSE Nov 9, 2017
README.md update for 0.3 ver Dec 12, 2017
setup.py glove Sep 3, 2017

README.md

DeepNLP-models-Pytorch

Pytorch implementations of various Deep NLP models in cs-224n(Stanford Univ: NLP with Deep Learning)

  • This is not for Pytorch beginners. If it is your first time to use Pytorch, I recommend these awesome tutorials.

  • If you're interested in DeepNLP, I strongly recommend you to work with this awesome lecture.

This material is not perfect but will help your study and research:) Please feel free to pull requests!!


Contents

Model Links
01. Skip-gram-Naive-Softmax [notebook / data / paper]
02. Skip-gram-Negative-Sampling [notebook / data / paper]
03. GloVe [notebook / data / paper]
04. Window-Classifier-for-NER [notebook / data / paper]
05. Neural-Dependancy-Parser [notebook / data / paper]
06. RNN-Language-Model [notebook / data / paper]
07. Neural-Machine-Translation-with-Attention [notebook / data / paper]
08. CNN-for-Text-Classification [notebook / data / paper]
09. Recursive-NN-for-Sentiment-Classification [notebook / data / paper]
10. Dynamic-Memory-Network-for-Question-Answering [notebook / data / paper]

Requirements

  • Python 3.5
  • Pytorch 0.2+
  • nltk 3.2.2
  • gensim 2.2.0
  • sklearn_crfsuite

Getting started

git clone https://github.com/DSKSD/cs-224n-Pytorch.git

prepare dataset

cd script
chmod u+x prepare_dataset.sh
./prepare_dataset.sh

docker env

ubuntu 16.04 python 3.5.2 with various of ML/DL packages including tensorflow, sklearn, pytorch

docker pull dsksd/deepstudy:0.2

pip3 install docker-compose
cd script
docker-compose up -d

cloud setting

not yet

References

Author

Sungdong Kim / @DSKSD