I have done this project with help of above article.
- There are 100 epochs to train the model
- The model consists of two models, one is Discriminator Model and other one is Generator Model.
- The Generator Model generates image4 and the Discrminatir Model discrinates whether it is real or fake.
- Once the discriminator model is unable to discriminate any more properly(50%), we can say that our generator model is generating very good image.
Explanation of the codes are commented inside notebook.
There are 25 images inside generated_images directory, which has been generated with our model.