You added automatic conversion from underscores to camelCase in 2.1 (#31), which is great but I think it would be better if it is optional.
I mean, if a property with underscore case found, parse the json node to that, and only try to parse to a camelCase one if a parameter with the exact same underscore case-d name not found.