The repository contains codebase for building state-of-the-art deep learning techniques for opinion mining.
Currenly following models have been created and tested:
- Models for IEST @ WASSA-2018 [FNN, CNN, LSTM, CNN-LSTM, LSTM-CNN ++]
- Related resources for featureizers (only if you are using them) (ex: lexicons, word-embedding models)
- Some python knowledge
- Install
textkit-learn - Configure path to resources in
oplab/config.py - Happy Experimenting!
-
To train a model its simple as calling
python oplab train -c [config_file] -m [model_path] -
To evaluate just
python oplab evaluate -m [model_path] -e [evaluation_metrics] -o [predictions_file]
- Code is under Apache License 2.0.
- Model (releases) are under Creative Commons Attribution - Non Commercial 2.0 Generic license.