Jaylon603
commented
1 year ago Generative Adversarial Network (GAN): GANs will be used for image generation of our virtual pets. GAN consists of two neural networks, a generator, and a discriminator, that compete against each other. The generator creates images, while the discriminator tries to distinguish between real and generated images. Training a GAN involves optimizing both networks until the generator creates realistic images.
Generator Training: The time complexity of training the generator in a GAN is dependent on the number of training iterations (epochs) and the batch size. If we were to train the generator for N epochs with a batch size of M, the time complexity would be roughly O(N * M), where N is the number of iterations and M is the batch size. Each iteration involves forward and backward passes through the generator network.
Discriminator Training: Training the discriminator would be a similar time complexity as the generator. The discriminator is trained for the same number of epochs as the generator with the same batch size. So, it also has a time complexity of O(N * M).
Data Preprocessing: Data preprocessing, including loading and transforming the dimensions of the training data(Pet Pixel Art), can have its time complexity depending on the dataset size and complexity of transformations applied. But for now we can assume time complexity would be O(N), where N is the size of the dataset.
API Integration: We also need to integrate the OpenAI API for the text generation on our website. We can assume time complexity would be O(1), since it would only need to be called once using a API key.
Dillon-G
Big - O analysis of overall System and Sub-Systems