Training convolutional network for classification tasks.
Based on "Convolutional Neural Networks for Sentence Classification" by Yoon Kim, link. Inspired by Denny Britz article "Implementing a CNN for Text Classification in TensorFlow", link.
Code is a fork of this implementation link
- Python3
- a focus on reusing pretrained word vectors
- a few code-decomposition-related modifications
- using argparse for remote configurable execution
- trying to implement configuration as close to the one in the paper as possible (comments are welcome)
- methods for working with datasets in a standard .csv (text, class_label) format
- L2 regularization
- ...
Code is easily portable to keras/Theano, Tensorflow-specific code is only in the GPU usage configuration part