Handle the "unsure" annotations

[thomashopkins32]

We also receive annotations that the experts who annotated the data are unsure about.

We could try the following:

• Treat them as full labels
• Treat them as full labels but with less of a penalty in the loss for getting them wrong (tunable)
• Smooth the labels to be 0.5 for these masks

Each of these options should be tested and examined in isolation on various sizes of the training set (probably need cross validation).

[thomashopkins32]

It looks like some annotations only have "unsure" as the label. This means that we should definitely be including it in our training somehow.