yma67 / Text-CNN

text-cnn

Tensorflow FULL Implementation of Yoon Kim's Text CNN Paper

Director: BOSS Li (Chris Li)

Experiment design and report: Shenshun Yao

Code: Yuxiang Ma

Requirement

1. numpy
2. gensim
3. tensorflow
4. tensorboard

Repeat of Experiment (Google Colab)

1. Specify Location of data files

to_trainer = TOTrainParameter(
    '/content/gdrive/My Drive/cs551final/data'
)

2. Make a dirctory "W2V" under "data", and put google w2v bin file under it.

3. Run utility functions and class models by "Runtime" -> "Run Before" on "Before Running your section, do a "Runtime" -> "Run before" on previous cells." cell.

4. choose your experiment to run.

Reference

Code Base was imported from https://github.com/dennybritz/cnn-text-classification-tf, with significant modifications to adapt the actual paper.