We try to detect sentence boundaries using deep learning. Created as part of the "Practical Applications of Multimedia Retrieval" seminar at the Hasso-Plattner-Institute, Potsdam, Germany.

We build a python-based demo using caffe.

#####Prerequirements:

1. Clone this repository
2. Install python 2.7 including the following packages from requirements.txt

pip install -r requirements.txt

3. Use the nltk downloader to download averaged_perceptron_tagger and punkt models:

python -m nltk.downloader

4. Setup caffe, like described here
5. Add path to the repository to your python path:

export PYTHONPATH=/path/to/sentence-boundary-detection-nn/python:$PYTHONPATH

6. Download Google Word Vector (GoogleNews-vectors-negative300.bin.gz) from here or use directly this url and extract the result into the sentence-boundary-detection-nn/python/demo_data directory
7. Paste your trained models into a demo data folder, for example sentence-boundary-detection-nn/python/demo_data with the following structure:

• lexical_models : containing all pretrained models you want to use in a seperate directory. Each models needs a
  • .ini
  • .caffemodel
  • net.prototxt file.
• text_data: containing all possible text files, which should be used as prediction input
• audio_models: containing all pretrainied audio models, each in a seperate directory. Each needs the same files as described for lexical models
• audio_examples: containing all audio files, which should be available during the demo. Each one in a seperate directory containing the ctm, energy and pitch files.

#####Start up

Change into the repository directory and execute, this should work right out of the box, unless you are using a custom demo_data folder:

python web_demo/web.py

Optionally you can specify the location of the word vector and the demo data. Otherwise default values are used. For further information execute:

python web_demo/web.py -h