/CS269_LEAM

Project for CS269_2019Spring

Primary LanguagePythonApache License 2.0Apache-2.0

CS269_LEAM

Project for CS269_2019Spring

How to run

see main.ipynb

Files

  • Since in our project, we try many different methods, some of which may use different libraries and apis. Instead of putting all of them into a uniform api. We split them by putting them to different directories.

  • The Code for purely LEAM:

    • data_process_reuters.ipynb : download and process reuters dataset.
    • main.ipynb : main file to run
    • src/
      • model.py :
        • Class Embedding : implement word embedding (pre-defined) along with label-embedding attention --- can change to bag of word or other predefined word embedding
        • Class Classifier : MLP + Relu classifier --- can change to LSTM / CNN / DNN
        • Class Leam_Classifier : combine two above
      • train.py : Code for training and testing
      • main.py : Main file to run in terminal
      • util.py : Util functions. --- please add all other function in util
      • zsdm.py : Please omit
    • data/ : will be produced by data_process_*.ipynb, please store every later data in this folder

  • The Code for ELMO:

    • The codes for ELMO is under the directory ELMO_BERT. The main interface is the ELMO.ipynb file. And this file use the functions in src directory
    • The code is based on google colabtory

Reference