Update README.md

README.md

@@ -1,7 +1,54 @@
 # Interpretable Adversarial Perturbation
-Code for Interpretable Adversarial Perturbation in Input Embedding Space for Text, IJCAI 2018.
+Code for [*Interpretable Adversarial Perturbation in Input Embedding Space for Text*](https://arxiv.org/abs/1805.02917), IJCAI 2018.
 
-# When the code will be published?
-- I am preparing the code of our paper now. 
-- The code will be pubished until June 23 or 24.
+This code reproduce the our paper with [Chainer](https://github.com/chainer/chainer).
 
+
+## Setup envirment
+Please install [Chainer](https://github.com/chainer/chainer) and [Cupy](https://github.com/cupy/cupy).
+
+You can set up the environment easily with this [*Setup.md*](https://github.com/aonotas/interpretable-adv/blob/master/Setup.md).
+
+## Download Pretrain Model
+Please download pre-trained model. Note that this pretrained model is genera
+```
+$ wget http://sato-motoki.com/research/vat/imdb_pretrained_lm_ijcai.model
+```
+
+
+# Run
+## Pretrain
+```
+$ python -u pretrain.py -g 0 --layer 1 --dataset imdb --bproplen 100 --batchsize 32 --out results_imdb_adaptive --adaptive-softmax
+```
+Note that this command takes about 30 hours with single GPU.
+
+## Train (VAT: Semi-supervised setting)
+```
+$ python train.py --gpu=0 --n_epoch=30 --batchsize 32 --save_name=imdb_model_vat --lower=0 --use_adv=0 --xi_var=5.0  --use_unlabled=1 --alpha=0.001 --alpha_decay=0.9998 --min_count=1 --ignore_unk=1 --pretrained_model imdb_pretrained_lm.model --use_exp_decay=1 --clip=5.0 --batchsize_semi 96 --use_semi_data 1
+```
+Note that this command takes about 8 hours with single GPU.
+
+## Train (Adversarial Training: Supervised setting)
+```
+$ python train.py --gpu=0 --n_epoch=30 --batchsize 32 --save_name=imdb_model_adv --lower=0 --use_adv=1 --xi_var=5.0  --use_unlabled=1 --alpha=0.001 --alpha_decay=0.9998 --min_count=1 --ignore_unk=1 --pretrained_model imdb_pretrained_lm.model --use_exp_decay=1 --clip=5.0
+```
+Note that this command takes about 6 hours with single GPU.
+
+# Authors
+We thank Takeru Miyato ([@takerum](https://github.com/takerum)) who suggested that we reproduce the result of a [Miyato et al., 2017].
+- Code author: [@aonotas](https://github.com/aonotas/)
+- Thanks for Adaptive Softmax implementation: [@soskek](https://github.com/soskek/)
+Adaptive Softmax: https://github.com/soskek/efficient_softmax
+# Reference
+```
+[Miyato et al., 2017]: Takeru Miyato, Andrew M. Dai and Ian Goodfellow
+Adversarial Training Methods for Semi-Supervised Text Classification.
+International Conference on Learning Representation (ICLR), 2017
+
+[Sato et al., 2018]: Motoki Sato, Jun Suzuki, Hiroyuki Shindo, Yuji Matsumoto
+Interpretable Adversarial Perturbation in Input Embedding Space for Text.
+IJCAI-ECAI-2018
+```
+# TODO
+- Add visualizing code