https://github.com/sonhm3029/neural-style-transfer-pytorch

This project is basic neural style transfer implement with pytorch
https://github.com/sonhm3029/neural-style-transfer-pytorch

Last synced: 3 months ago
JSON representation

This project is basic neural style transfer implement with pytorch

• Host: GitHub
• URL: https://github.com/sonhm3029/neural-style-transfer-pytorch
• Owner: sonhm3029
• Created: 2023-03-26T10:16:53.000Z (about 2 years ago)
• Default Branch: master
• Last Pushed: 2023-09-04T08:48:48.000Z (almost 2 years ago)
• Last Synced: 2025-01-17T01:24:08.763Z (5 months ago)
• Language: Jupyter Notebook
• Size: 12.8 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Neural-style-transfer-pytorch

## Overview

### 1. Build model to extract features

- Using pretrained VGG19 model

- Using only layer index 0, 5, 10, 19, 28

### 2. Load images

- Load content image

- Load style image

- Generate random white noise target image

=> Resize style and target image to size of content image

![](./src/assets/1.png)

### 3. Init hyperparameters and optimizer:

- style_weight: 1000,

- content_weight: 1,

- learning_rate: 0.001

- epochs: 6000

Using Adam optimizer with parameters is target_images

### 4. Training:

- Calculate content loss

$$content\_loss = \frac{1}{2}(content\_img - target\_img)^2$$

- Calculate style loss:

Gram matrix:

$$G = \frac{1}{c*h*w}(target\_img.target\_img^T)$$

$$A = \frac{1}{c*h*w}(style\_img.style\_img^T)$$

$$style\_loss = \frac{1}{2*c*h*w}.(G - A)^2$$