In this homework, you have to select to implement a model among the list below and implement a code to train a set of data. The dataset should be data that has never been used before in the class (cannot be iris, mnist, cifar10, glass. If you choose reinforcement learning, use another gym (not LunarLander).

The choices of the model are

1. Autoencoder (normal or VAE)

2. GANs

3. Deep reinforcement learning model (deep-Q, policy-gradient, or Actor-critic)

What to submit: (in a single pdf file)

1. Description of the problem and data that you are trying to solve.

2. Model architecture such as model.summary()

3. Printed Source code

4. Graph Results that indicate learning progress (e.g. loss). For GAN, you can use output samples to show the progress.

5. Final results (if applicable).

Can you guide me the framework and data in one of the 3 models?