Pytorch implementation of GAN
This is result at 0 Epoch on batch 200. We can see that the Generative model is learning something.
After 210 Epoch at batch 200. We can see that the Generative model has learned some weights to recreate legible numbers.
After almsot 800 epochs later, at 1000. The Generative model has learned more sophisticated features. GAN are known for its difficulty of converging but our results are good enough.
-- You can check the images folder for some results
Instruction: If ever you want to run this code. Just clone the repo. Delete the folder called images. The code will recreate the folder structures.