This repository contains the iPython notebook and training data to accompany the O'Reilly tutorial on sentiment analysis with LSTMs in Tensorflow. See the original tutorial to run this code in a pre-built environment on O'Reilly's servers with cell-by-cell guidance, or run these files on your own machine. There is also another file called
Pre-Trained LSTM.ipynb which allows you to input your own text, and see the output of the trained network.
Before running the notebook, you'll first need to download all data we'll be using. This data is located in the
training_data.tar.gz tarballs. We will extract these into the same directory as
Oriole LSTM.ipynb. As always, the first step is to clone the repository.
git clone https://github.com/adeshpande3/LSTM-Sentiment-Analysis.git
Next, we will navigate to the newly created directory and run the following commands.
tar -xvzf models.tar.gz tar -xvzf training_data.tar.gz
In order to run the iPython notebook, you'll need the following libraries.
In order to load the models without errors you need to convert the checkpoints using the converter provided by TensorFlow:
wget https://raw.githubusercontent.com/tensorflow/tensorflow/master/tensorflow/contrib/rnn/python/tools/checkpoint_convert.py python checkpoint_convert.py models/pretrained_lstm.ckpt-90000 converted-checkpoints/pretrained_lstm-90000.ckpt
You should also replace the original models folder if you don't want to modify the code:
rm -rf models mv converted-checkpoints models
With Docker, you could just mount the repository and exec it.
Install Docker. Follow the docker guide.
Build docker image
cd LSTM-Sentiment-Analysis docker build -t="@yourname/tensorflow_1.1.0_py3" .
Run the container from the image
docker run -p 8888:8888 --name=tensorflow_yourname_py3 -v /@YourDir/LSTM-Sentiment-Analysis:/LSTM-Sentiment-Analysis -it @yourname/tensorflow_1.1.0_py3
and visit the URL(http://localhost:8888/)
Stop and restart the container
docker stop tensorflow_yourname_py3 docker start tensorflow_yourname_py3 docker attach tensorflow_yourname_py3
If jupyter is down, relaunch it by using the command below.
cd LSTM-Sentiment-Analysis jupyter notebook --ip=0.0.0.0 --allow-root
The easiest way to install TensorFlow as well as NumPy, Jupyter, and matplotlib is to start with the Anaconda Python distribution.
Follow the installation instructions for Anaconda Python. We recommend using Python 3.6.
Follow the platform-specific TensorFlow installation instructions. Be sure to follow the "Installing with Anaconda" process, and create a Conda environment named
If you aren't still inside your Conda TensorFlow environment, enter it by opening your terminal and typing
source activate tensorflow
If you haven't done so already, download and unzip this entire repository from GitHub, either interactively, or by entering
git clone https://github.com/adeshpande3/LSTM-Sentiment-Analysis
cd to navigate into the top directory of the repo on your machine
Launch Jupyter by entering
and, using your browser, navigate to the URL shown in the terminal output (usually http://localhost:8888/)