Online sales of antique and unique art were reported at $6 billion and growing by 11% each year. For Upper middle class families to medium scale business, it can be very tricky to buy exclusive and unique art.
To offer unique and exclusive art at a very modest price, I have developed a web application which offers machine generated abstract landscape art which you can buy for $2. The art gallery refreshes every 15 minutes, the art can be yours or lose it forever.
https://drive.google.com/drive/u/1/folders/1az6J5ucolVTDz3x3ktcrNeIYuoErPwZi
Quick view of progressive Training.
There are two classes landscape and abstract on which the GAN model is trained. Approximately GAN model requires 3000 images to completely learn one class.
The progressive growing of GANs trains the GAN network in multiple phases. In phase 1, it takes in a latent feature z and uses two convolution layers to generate 4×4 images. Then, we train the discriminator with the generated images and the 4×4 real images. Once the training stables, we add 2 more convolution layers to upsampling the image to 8×8 and 2 more convolution layers to downsampling images in the discriminator.
I have 6 phases in total, to generate the 128 x 128 abstract landscape images
GAN_GEN_SHADOW_5.pth and GAN_GEN_5.pth are two pre trained generator network which you can directly use to generate various types of images, including 128x128 landscape oil paintings.
Import pro_gan generator into your code as shown in GanScapeStudio.py and load .pth trained weights into the generator and generate abstract landscape images.
To train on your custom dataset using Progressive Growing of GANs, there is an example in Notebook/trainpgGAN.ipynb Set hyperparameteres like depth of training model, number of epochs, fade ins, batch size and feed back factor.
ProGAN- https://arxiv.org/abs/1710.10196
Animesh Karewar Blog- https://medium.com/@animeshsk3/the-unprecedented-effectiveness-of-progressive-growing-of-gans-37475c88afa3
DCGAN PyTorch- https://pytorch.org/tutorials/beginner/dcgan_faces_tutorial.html
Wiki- https://www.wikiart.org/en/paintings-by-genre/landscape?select=featured#!#filterName:featured,viewType:masonry
This repository is only for educational purpose created during Insight Data Science fellowship.