/Indian_Name

An implementation of character based language model to predict Indian names

Primary LanguagePythonMIT LicenseMIT

Indian_Name

This repository is divided into four parts:

  1. data - contains the cleaned and uncleaned data in one of four files:

    1. Name.txt, Name.csv - contains the data after being cleaned using cleaner program '../coursera_based_implementation/name_cleaner'
    2. Indian_Male_Names.csv, Indian_Female_Names.csv - contains the database downloaded from kaggle at 'https://www.kaggle.com/chaitanyapatil7/indian-names/version/1'

  2. keras_based_implementation - contains the keras with tensorflow based implementation for the model. Using SimpleRNN

    1. name_predicter.py - program to predict the name given a trained model and parse the data to the trainer
    2. name_train.py - program to train the neural network on dataset

  3. coursera_based_implementation - contains the implementation based on the coursera course on sequence models by Andrew Ng

    1. utils.py - contains the utilities to make the 'name_pred.py' predicter work
    2. name_pred.py - contains the predicter program
    3. name_cleaner.py - cleans the data and extracts only the first name from the data files and saves in Name.txt, Name.csv

  4. models - contains keras models trained on different number of epochs