Skip to content
Neural Message Passing for Multi-Label Classification
Branch: master
Clone or download
Latest commit ee95b43 May 9, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data First commit Apr 5, 2019
lamp name change Apr 5, 2019
models name changes Apr 10, 2019
visualizations First commit Apr 5, 2019
DataLoader.py First commit Apr 5, 2019
LICENSE Initial commit Apr 5, 2019
README.md Update README.md May 9, 2019
config_args.py name changes Apr 10, 2019
data_loader.py First commit Apr 5, 2019
evals.py First commit Apr 5, 2019
get_best_threshold.py First commit Apr 5, 2019
preprocess.py First commit Apr 5, 2019
train.py add testing only feature Apr 10, 2019
utils.py First commit Apr 5, 2019
visualize.py First commit Apr 5, 2019

README.md

"Neural Message Passing for Multi-Label Classification"
Jack Lanchantin, Arshdeep Sekhon, Yanjun Qi
https://arxiv.org/abs/1904.08049

Requirement

  • python 3.4+
  • pytorch 0.2.0
  • tqdm
  • numpy

Usage

Download the data

http://www.cs.virginia.edu/~jjl5sw/data/lamp_datasets.tar.gz (745M)

Train and Test the model (training script contains the validation and testing code)

python train.py -dataset reuters -batch_size 32 -d_model 512 -d_inner_hid 512 -n_layers_enc 2 -n_layers_dec 2 -n_head 4 -epoch 50 -dropout 0.2 -dec_dropout 0.2 -lr 0.0002 -encoder 'graph' -decoder 'graph' -label_mask 'prior'

Acknowledgement

Much of this code was adapted from @jadore801120

You can’t perform that action at this time.