How to Create A Word Generator with Tensorflow in Python - Our RNN
In this series, I’m experimenting with Natural Language Processing, specifically with Word Generation using Python and Tensorflow! Watch as we write the functions to process the input (In this case, a bunch of text files that have all written works by Edgar Allan Poe) and then train a model to predict a “next character” given a sequence of characters.
Tensorflow provides some really fun tools that can be used to do all sorts of interesting programming tasks, and this is just a small sample of it’s power. But I think it’s cool to learn and hopefully this series gives you a good, small foundation for python and tensorflow.
🔴 Link to Episode 1!
• How to Create a Word Generator with T...
🔴 Check out Programming With Chris Live T/Th/Sat nights!
https://bit.ly/3nHuLuN
🔴 Subscribe For More Programming Tutorials
https://bit.ly/3NN23DB
🔴 Download Source Code used in the Series Here!
https://bit.ly/3AtSKFu
In this video, I go over how to process the text input, how we build a Tensorflow model with the keras package, how we convert that text into numbers that the model can use to train itself, and then how we use a built model to actually generate text! The model used in this series is a Recurrent Neural Network (RNN) with two LSTM layers, with ReLu activation functions.
🕘Timestamps:
0:30 Covering our functions from last time
05:39 - Talking about what we're doing in this video
07:55 - How to build a function that trains our RNN model
35:30 - Fixing our model input mistake (Used the full character array instead of the character vocabulary)
54:50 - How to build a function that generates text from our RNN
01:06:10 - Outro
Thanks for watching! I plan on making a ton more tutorial programming videos on libraries I find interesting, and if that’s your sort of thing, check out my other videos and subscribe for more. Thanks!
🎬Watch My most recent videos:
https://bit.ly/3yJwaY5