Skip to content
Minecraft texture pack creator using a conditional deep convolutional generative adversarial network
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data-minecraft/train
example-texture-pack
output-test
texture-pack-extraction
.gitignore
CDCGAN-generator-test1600.h5
README.md
main.py

README.md

CDCGAN

Minecraft texture pack creator using a conditional deep convolutional generative adversarial network. Made in 5 weeks for a university module.

30-second trailer

https://www.youtube.com/watch?v=NBEB2Dkw1aE

Weekly Development Videos

https://www.youtube.com/watch?v=Ur6vrh5-gMU&list=PLAu3dU8p746BB1z4TUxNWb8e1J3OjnxYR&index=11

Samples - 1600 Epochs

Output

Discriminator Loss

Discriminator Accuracy

Generator Loss

Sample Input

Instructions

main.py is the CDCGAN and it has a few command line options:

If you do not specify an option, it will train the CDCGAN from scratch.

If you specify "all", it will output 100 of every texture using CDCGAN-generator-test1600.h5

If you specify "minecraft", it will generate a texture pack using CDCGAN-generator-test1600.h5

If you specify a specific texture's label, it will generate that texture using CDCGAN-generator-test1600.h5

There's a sample texture pack, and within "texture-pack-extraction" you will find the extraction process for getting training data and some existing data.

You can’t perform that action at this time.