Adding Broken-Up Files for mnist.py

[e-corwin]

• Created new files that separate the functions of the mnist model: initializing.py, optimizing.py, automatic_differentiation.py, saving_and_loading.py
• [initializing.py] initializes the networks layers by applying transformations on the image inputs
• [optimizing.py]: trains and tests the model on FashionMNIST data, optimizes the model based on the training, sets hyperparameters
• [automatic_differentiation.py]: backpropagation? not sure if/why this would be separate from training/optimizing
• [saving_and_loading.py]: saves the model to a path for loading and running

[siddk]

So typically when writing PRs, we should use the description box to provide a concise, bulleted list of changes; can you edit the box above @e-corwin?

[siddk]

Other high-level comment - in Python, the normal format for filenames/directory names is to use all lowercase names with underscores (_) separating words.

This is called "snake_case", and is just a different convention from Java's "CamelCase"

[siddk]

So this is a good linear split, but you see how we're duplicating code across files? How do we avoid that, and actually "import" code blocks from other files?

Let's talk about this today!

[siddk]

Similarly, saving and loading should be functions of the existing models we've trained!

[e-corwin]

So typically when writing PRs, we should use the description box to provide a concise, bulleted list of changes; can you edit the box above @e-corwin?

By PR I assumed you were referring to the comment box, I updated the box with more specificity to how I separated my files!