Name Generator Neural Network Using Pytorch

This repository contains a neural network-based models for generating unique names. The model is trained on a dataset of names provided in the names.txt file. The complete implementation dont use any inbuilt libraries or methods ,but only basic modules like numpy,Pytorch,matplotlib.

Project Structure

• names.txt: Corpus of names used for training the models.

• 1_Bigram_Model.ipynb: Notebook implementing a bigram model. It counts frequencies of character pairs and generates names based on probabilities.

• 2_Bigram_Neural_Net.ipynb: Notebook implementing a bigram neural network. It uses the previous character to generate the next character in a name.

• 3_Ngram_Neural_Net.ipynb: Notebook implementing an N-gram neural network. It uses the previous N characters to generate the next character in a name.

Usage

1. Dataset: Ensure that the names.txt file contains the corpus of names you want to train on.

2. Bigram Model: Open and run the 1_Bigram_Model.ipynb notebook to see the implementation of the bigram model.

3. Bigram Neural Net: Open and run the 2_Bigram_Neural_Net.ipynb notebook to see the implementation of the bigram neural network.

4. N-gram Neural Net: Open and run the 3_Ngram_Neural_Net.ipynb notebook to see the implementation of the N-gram neural network.