Containers

For convenience, MiniML provides containers to store multiple parameters and models within a bigger one:

• MiniMLParamList for parameters,
• MiniMLModelList for models.

You should use these whenever you want to pack multiple parameters or models inside of a bigger model. If you simply add as child of a MiniMLModel a bare list, it won't get properly bound to the buffer and the model won't work.

class MyModel(MiniMLModel):
    def __init__(self):
        self._layers = [Linear(3,5), Activation(), Linear(5,1)] # WRONG
        self._layers = MiniMLModelList([Linear(3,5), 
                                        Activation(), 
                                        Linear(5,1)]) # Correct