Skip to content

Latest commit

 

History

History
executable file
·
42 lines (32 loc) · 750 Bytes

Readme.md

File metadata and controls

executable file
·
42 lines (32 loc) · 750 Bytes
Raw

DeepCoNN

This is the implement for NARRE, originated from https://github.com/chechongthu/NARRE

Environments

  • python >= 3.5.0
  • Tensorflow >= 1.4.0
  • numpy
  • pandas
  • gensim

1. Data Preprocessing

Preprocess data for training model

1.1 split data as train set, valid set and test set;

python3 loaddata.py

1.2 get pre-trained word embedding

python3 vocab_and_emb.py

1.3 get hyperparameters of dataset

python3 data_pro.py

2.Training

train model

python3 train.py

Main directories and files

  • ./data: storage for datasets and embeddings
  • ./data_pro: codes for preprocessing data
  • ./src/train.py: codes for training model
  • ./src/model.py: codes for building model