def __init__(
self,
freq: str,
context_length: int,
prediction_length: int,
trainer: Trainer = Trainer(),
num_layers: int = 1,
num_cells: int = 50,
cell_type: str = "lstm",
num_eval_samples: int = 100,
cardinality: List[int] = list([1]),
embedding_dimension: int = 10,
distr_output: DistributionOutput = StudentTOutput(),
) -> None:
model = RNN(mode=cell_type,
num_layers=num_layers,
num_hidden=num_cells)
super(CanonicalRNNEstimator, self).__init__(
model=model,
is_sequential=True,
freq=freq,
context_length=context_length,
prediction_length=prediction_length,
trainer=trainer,
num_eval_samples=num_eval_samples,
cardinality=cardinality,
embedding_dimension=embedding_dimension,
distr_output=distr_output,
)
def __init__(
self,
mode: str,
hidden_size: int,
num_layers: int,
bidirectional: bool,
**kwargs,
) -> None:
assert num_layers > 0, "`num_layers` value must be greater than zero"
assert hidden_size > 0, "`hidden_size` value must be greater than zero"
super().__init__(**kwargs)
with self.name_scope():
self.rnn = RNN(mode, hidden_size, num_layers, bidirectional)
def __init__(
self,
mode,
num_hidden,
num_layers,
num_output,
bidirectional=False,
**kwargs,
):
super(RNNModel, self).__init__(**kwargs)
self.num_output = num_output
with self.name_scope():
self.rnn = RNN(
mode=mode,
num_hidden=num_hidden,
num_layers=num_layers,
bidirectional=bidirectional,
)
self.decoder = nn.Dense(
num_output, in_units=num_hidden, flatten=False
)
def __init__(
self,
mode: str,
hidden_size: int,
num_layers: int,
bidirectional: bool,
use_static_feat: bool = False,
use_dynamic_feat: bool = False,
**kwargs,
) -> None:
assert num_layers > 0, "`num_layers` value must be greater than zero"
assert hidden_size > 0, "`hidden_size` value must be greater than zero"
super().__init__(**kwargs)
self.mode = mode
self.hidden_size = hidden_size
self.num_layers = num_layers
self.bidirectional = bidirectional
self.use_static_feat = use_static_feat
self.use_dynamic_feat = use_dynamic_feat
with self.name_scope():
self.rnn = RNN(mode, hidden_size, num_layers, bidirectional)