naamkaran: Generative Model for Names

Tests PyPI Version Downloads

Naamkaran is a generative model for names built with PyTorch. It uses a character-level RNN (LSTM) trained on Florida Voter Registration Data to generate names based on starting letter, ending letter, gender, and other parameters.

Features

  • Character-level LSTM: Deep learning model trained on real name data

  • Flexible generation: Generate names by starting letter, ending letter, gender, and length

  • Multiple interfaces: Python API, command-line tool, Gradio web app, and Flask API

  • High-quality output: Names that look and sound realistic

  • Fast inference: Optimized for quick name generation

Quick Start

Install naamkaran:

pip install naamkaran

Generate names programmatically:

from naamkaran.generate import generate_names

# Generate 5 female names starting with 'A' and ending with 'a'
names = generate_names(
    starting_letter='A',
    ending_letter='a',
    gender='F',
    num_names=5,
    max_len=8,
    temperature=0.7
)
print(names)

Or use the command line:

generate_names -s A -e a -g F -n 5 -m 8 -t 0.7

Table of Contents

User Guide:

API Reference:

Development:

Indices and Tables