Optimizer.init(parameters: dict)#

Initialize the optimizer’s state

This function can be used to initialize optimizers which have state (like momentum in SGD). Using this method is optional as the optimizer will initialize itself if the state is not yet set. However, there are some cases where explicit initialization is useful in order to have access to the Optimizer.state before the first call to Optimizer.update().


model (dict) – A Python tree of parameters.


>>> optimizer = optim.SGD(learning_rate=1e-1, momentum=0.9)
>>> model = nn.Linear(2, 2)
>>> optimizer.init(model.trainable_parameters())
>>> optimizer.state.keys()
dict_keys(['step', 'learning_rate', 'weight', 'bias'])