Requirements

You may be able to use the library without this requirements, but the library is tested only from the below environment.

• One or more GeForce GTX 1080 Ti graphics cards (driver should be installed)
• CentOS 7

Installation

1. Install bzip2

sudo yum install bzip2

Follow instructions.

2. Install Anaconda with Python 3.6

cd ~
mkdir Downloads
cd ~/Downloads
wget https://repo.continuum.io/archive/Anaconda3-5.1.0-Linux-x86_64.sh
bash Anaconda3-5.1.0-Linux-x86_64.sh

Follow instructions.
Answer 'yes' to prepend the Anaconda3 install location to PATH in your .bashrc.
Answer 'no' not to install Microsoft VSCode.

3. Install NVIDIA CUDA Toolkit 9.1

wget https://developer.nvidia.com/compute/cuda/9.1/Prod/local_installers/cuda_9.1.85_387.26_linux
wget https://developer.nvidia.com/compute/cuda/9.1/Prod/patches/1/cuda_9.1.85.1_linux
wget https://developer.nvidia.com/compute/cuda/9.1/Prod/patches/2/cuda_9.1.85.2_linux
wget https://developer.nvidia.com/compute/cuda/9.1/Prod/patches/3/cuda_9.1.85.3_linux
sudo sh cuda_9.1.85_387.26_linux

Follow instructions.
Do not install NVIDIA Accelerated Graphics Driver, if you have already installed it.

sudo sh cuda_9.1.85.1_linux
sudo sh cuda_9.1.85.2_linux
sudo sh cuda_9.1.85.3_linux

Follow instructions.

4. Install PyTorch

conda install pytorch torchvision cuda91 -c pytorch

Follow instructions.

5. Install ParlAI

python setup.py develop

6. Copy files

• Bot
  cc/exp-ko_multi_20180316/dict_file_100000.dict
  cc/exp-opensub_ko_nlg/dict_file_100000.dict
  cc/exp/exp-emb200-hs2048-lr0.0001-multi2018_30000/exp-emb200-hs2048-lr0.0001-multi2018_30000
  cc/exp/exp-emb200-hs1024-lr0.0001-oknlg/exp-emb200-hs1024-lr0.0001-oknlg
  data/word2vec_ko/ko.bin
• Emotional Bot
  cc/exp-opensub_kemo_all/dict_file_100000.dict
  cc/exp-opensub_ko_nlg/dict_file_100000.dict
  cc/exp/exp-emb200-hs2048-lr0.0001-allK/exp-emb200-hs2048-lr0.0001-allK
  cc/exp/exp-emb200-hs1024-lr0.0001-oknlg/exp-emb200-hs1024-lr0.0001-oknlg
  data/word2vec_ko/ko.bin
• Context-Aware Emotional Bot
  cc/exp-opensub_kemo_20190226/dict_file_100000.dict
  cc/exp-opensub_ko_nlg/dict_file_100000.dict
  cc/exp/exp-emb200-hs4096-lr0.0001-emotional_hred/exp-emb200-hs4096-lr0.0001-emotional_hred
  cc/exp/exp-emb200-hs1024-lr0.0001-oknlg/exp-emb200-hs1024-lr0.0001-oknlg
  data/word2vec_ko/ko.bin

7. Install KoNLPy

sudo yum install gcc-c++ java-1.7.0-openjdk-devel python-devel
pip install JPype1-py3
pip install konlpy

8. Install Gensim

pip install gensim

Usage

• Bot

from examples.bot import Bot

bot = Bot('exp/exp-emb200-hs2048-lr0.0001-multi2018_30000/exp-emb200-hs2048-lr0.0001-multi2018_30000', 'exp-ko_multi_20180316/dict_file_100000.dict', True)
answer = bot.reply('안녕')

• Emotional Bot

from examples.bot import Bot

bot = Bot('exp/exp-emb200-hs2048-lr0.0001-allK/exp-emb200-hs2048-lr0.0001-allK', 'exp-opensub_kemo_all/dict_file_100000.dict', True)
answer, emotion = bot.reply('안녕', 'Neutral') # The second parameter can be one of these: Neutral, Happiness, Anger, Sadness, Surprise, Fear, Disgust.

• Context-Aware Emotional Bot

from examples.bot import Bot

bot = Bot('exp/exp-emb200-hs4096-lr0.0001-emotional_hred/exp-emb200-hs4096-lr0.0001-emotional_hred', 'exp-opensub_kemo_20190226/dict_file_100000.dict', True)
answer, emotion = bot.reply('안녕', 'Neutral', 'test0') # The second parameter can be one of these: Neutral, Happiness, Anger, Sadness, Surprise, Fear, Disgust.

The last parameter of the Bot constructor determines whether to use CUDA or not.
False is not tested.