Generate TV scripts using Recurrent Neural Network
In this project a new type of Neural Networks is used to generate fake TV scripts of a TV show and it is one of the most interesting project i have done in Deep learning Nano-Degree using Recurrent neural networks. In this project, I have generated my own Simpsons TV scripts using RNNs. I have used part of the Simpsons dataset of scripts from many seasons. The Neural Network I have built will generate a new, "fake" TV script. and in this project many concepts are covered like Recurrent Neural Networks, Tensorflow, Embeddings and Word2Vec. and there are severla steps for doing this job:
- Get and explore the Data
- preprocess the data
- create Lookup Table, give tokens to the words
- save the data
- Build the Neural Network
- modeling and training the network
- tuning the hyperparameters
- Generate TV Script
- Generate Text and new script
this code is written in python language and some libraries are needed: Python 3, Numpy, Pandas, Matplotlib, Torchvision, PyTorch, TensorFlow, Keras