/tv-script-generation-using-RNN

Project from Deep Learning Nanodegree - Udacity

Primary LanguageJupyter NotebookOtherNOASSERTION

TV Script Generator

Project Overview

In this project, I trained an RNN model on Seinfeld TV scripts. After the training, the Neural Network was able to generate "fake" TV script, based on patterns it reconginzes in this training data. The script was more or less readbale. you could find the same at the end of the jupyter notebook

Project Instruction

Instruction

  1. Clone the repository and navigate to the downloaded folder. 
    	git clone https://github.com/eswar3/tv-script-generation-using-RNN.git
	cd tv-script-generation-using-RNN
  2. Open the tv_script_generation.ipynb file. 
    	jupyter notebook tv_script_generation.ipynb
  3. Read and follow the instructions in the notebook

Project Steps

  • Explore the Data and Implement Pre-processing Functions
  • Build the Neural Network
    • Define forward and backpropagation
  • Neural Network Training
    • Train Loop
    • Hyperparameters
    • Train
  • Generate TV Script
    • Generate a new script

Libraries