zhongyunuestc/lda_bert

LDA/BERT Topic Modeler

Overview

This code is modified from: https://www.kaggle.com/dskswu/topic-modeling-bert-lda

Prepare Workspace

• Create virtual environment (called ptorch):
  • run conda env create --file ldabert.yml
  • conda activate ptorch
• Install additional dependencies:
  • start python and run:

    import nltk
    nltk.download('averaged_perceptron_tagger')
    nltk.download('stopwords')
    nltk.download('punkt')
    exit()
    

• Run Jupyter From Within venv:
  • jupyter notebook
  • Open up /examples/bert_lda_tm.ipynb