Configuring partial functions

[JIy3AHKO]

In some scenarios I want the result of cfn.Config, to be a partial function. E. g.:

adamw = cfn.Config(
   functools.partial,
   torch.optim.AdamW, 
   lr=1e-3,
   weight_decay=0.01
) 

@cfn.config(optimizer=adamw)
def train(optimizer: Callable):
    model = Model()
    opt = optimizer(model.params())
    ...

Should we make some shortcut for such scenarios? Since with this implementation it wouldn't be possible to change the class of adamw (we have no support for *args override yet). The shortcut could look like:

optimizer = cfn.partial_config(torch.optim.AdamW, lr=1e-3, weight_decay=0.01)

@cfn.config(optimizer=optimizer)
def train(optimizer: Callable):
    model = Model()
    opt = optimizer(model.params())
    ...