Tensorflow Object Detection Tutorial: Fast Style Transfer in Real-time
#tensorflow #styletransfer #tensorflowhub #python
In this episode, I will show you how to make really cool style transfer videos. You can do this either on your computer if you have an Nvidia GPU or on Google Colab, which gives you free access to a GPU.
To make my style transfer videos, I will be using some famous paintings like Starry Night from Van Gogh, the Munch Scream by Edvard Munch, just to name a couple.
Fast Style Transfer
Fast Style Transfer is a technique in which given a content image and a style image, you apply the style image to the content image and you get as a result a new stylized image that incorporates the style of the given style image. Simple right?
Tensorflow Hub
For this tutorial, I will be using Tensorflow 2 and a pre-trained style transfer Deep Neural Network model from the Tensorflow Hub.
The Tensorflow Hub is a repository of pre-trained models that you can fine-tune and you can deploy immediately, with a minimum amount of code and effort.
Without further ado, let’s get started!
FOLLOW ME
Instagram: https://www.instagram.com/codemental/
Blog: https://spltech.co.uk/blog/
SUPPORT THE CHANNEL or get MENTORING
http://www.patreon.com/codemental
RESOURCES
Tensorflow 2 Style Transfer Tutorial – https://www.tensorflow.org/hub/tutorials/tf2_arbitrary_image_stylization
SOURCE CODE:
https://github.com/armindocachada/fast-style-transfer-tensorflowjs/blob/main/style_transfer.ipynb
RELATED VIDEOS
How to run any Tensorflow model on a browser with Tensorflow.js Easy!
A demo of Real-time Fast Style Transfer model with Tensorflow.js and REACT/Next.js in the Browser
TIMESTAMPS
0:00 Introduction
1:00 Starting Anaconda
source