A conversational Chatbot with tensorflow
Dependencies
-numpy
-scipy
-six
-tensorflow 0.12.1
Use pip to install any missing dependencies
Usage
1.) To train the bot, edit seq2qeq.ini file so that mode is set like
mode = train
2.) Then run the code like
python chat.py
3.) To test the bot after training, edit seq2seq.ini file so that mode is set like
mode = test
4.) then run the code like
python chat.py
This cornell movie corpus contains a large metadata-rich collection of fictional conversations extracted from raw movie scripts:
1.) Download the processed data by running pull_data.sh
./pull_data.sh
2.) After runninh pulldata script following files will be generated:
train.enc
train.dec
test.enc
test.dec
3.) Put the above generated files in /data folder