Branch: master
Find file History
giuseppefutia and ylfdq1118 [Tutorial] Update prerequisites of README (#380)
* Update prerequisites of README

* dependencies for pytorch models

* dependencies for mxnet models

* minor
Latest commit 6c3dba8 Feb 11, 2019
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
README.md [Tutorial] Update prerequisites of README (#380) Feb 12, 2019
train.py [Model][MXNet] MXNet Tree LSTM example (#279) Dec 27, 2018
tree_lstm.py [model] Tree-LSTM update & hotfix(dropbox link) (#202) Dec 1, 2018

README.md

Tree-LSTM

This is a re-implementation of the following paper:

Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks Kai Sheng Tai, Richard Socher, and Christopher Manning.

The provided implementation can achieve a test accuracy of 51.72 which is comparable with the result reported in the original paper: 51.0(±0.5).

Data

The script will download the [SST dataset] (http://nlp.stanford.edu/sentiment/index.html) automatically, and you need to download the GloVe word vectors yourself. For the command line, you can use this.

wget http://nlp.stanford.edu/data/glove.840B.300d.zip
unzip glove.840B.300d.zip

Dependencies

  • PyTorch 0.4.1+
  • requests
  • nltk
pip install torch requests nltk

Usage

python train.py --gpu 0

Speed

On AWS p3.2x instance, it can achieve 3.18s per epoch when setting batch size to 256.