Skip to content

Latest commit

 

History

History
96 lines (76 loc) · 2.86 KB

README.md

File metadata and controls

96 lines (76 loc) · 2.86 KB

AttentionedDeepPaint

Automatic Anime Sketch Colorization Using Generative Adversarial Network (GAN)

Dependency

Refer requirements

Install

  1. Pytorch (>= 0.4.1) (No problem in torch 1.0.1)
  2. Torchvision (>= 0.2.1) based on your python version, os, cuda version etc...

Usage

Download dataset

  1. go to link and download
  2. unzip

Train

$ ./train.sh

Reference

  1. Sketch Image Generation
  1. Color Histogram Extraction
  1. Attentioned UNet

Colorize Result Examples

refer Result1 Result2 Result3 Result4 Result5 Result6 Result7 Result8 Result9 Result10 Result11 Result12 Result13 Result14 Result15 Result16 Result17

To see experiment trials, click

General Result Example (Same Sketch, Different Styles)

Gen1 Gen2 Gen3 Gen4 Gen5 Gen6 Gen7 Gen8 Gen9 Gen10 Gen11 Gen12

Attention Map Result Examples

attrefer Att1 Att2 Att3 Att4 Att5

Train Settings & Log

Hyperparameter Value
learning_rate 0.0002
beta1 (optimizer) 0.5
batch_size 4
lambda 100
epochs 30
learning_rate schedule None
Discriminator PatchGAN
Weight Initialization (0, 0.02) Normal Distribution

Log