Implement Flatten.forward() [10 points]

[chris-dare]

In nn/conv.py, complete Flatten.forward(). Note that this is in the conv.py file, not in functional.py.

This layer is often used between Conv and Linear layers, in order to squish the high-dim convolutional outputs into a lower-dim shape for the linear layer. For more info, see the torch documentation and the example we provided in the code comments.

Hint: This can be done in one line of code, with no new operations or (horrible, evil) broadcasting needed.

Bigger Hint: Flattening is a subcase of reshaping. np.prod() may be useful.