Morphology Segmentation using DPGMM
Research project done as a part of my research internship at NAAMII.
- Add a raw Nepali text corpus in the
data
folder. You can select the name of the file in run.py - Set the parameters in run.py
- Run run.py
- After every 10 iterations of the Gibbs Sampling, model is saved in
models
directory. - You can continue the training later with a saved model.
- After every 10 iterations of the Gibbs Sampling, model is saved in
- Run inference.py for inference
All the mathematical background can be found here.