/Neural-Styling

Primary LanguagePython

Neural Style Transfer!

Implemented a Neural Style Transfer using Tensorflow, Keras and StreamLit. This project shows how a a neural network transforms the source image to target image at multiple iterations.

Steps

  1. User must select the target image and a source image like a painting, design etc.
  2. Then select the epoch for how many times the model should run and at what iterations the converted image must be shown. And then click "Start" button.

Requirements

TensorFlow = 2.1.0 Keras = 2.3.1 StreamLit = 0.70.0

To download use: pip install requirement.txt

Source & Target Image

Source Image Target Image
Source Image Target Image

Images Generated At Different Epochs

Image At Iteration 0
Image At Iteration 0
Image At Iteration 100
Image At Iteration 100
Image At Iteration 200
Image At Iteration 200
Image At Iteration 300
Image At Iteration 300
Image At Iteration 400
Image At Iteration 400
Image At Iteration 500
Image At Iteration 500
Image At Iteration 600
Image At Iteration 600
Image At Iteration 700
Image At Iteration 700