Sequence1dBlock

class deeplay.blocks.sequence.sequence1d.Sequence1dBlock(*args, **kwargs)

Bases: BaseBlock

Convolutional block with optional normalization and activation.

Attributes Summary

is_recurrent

Methods Summary

GRU()
LSTM()
RNN()
append_dropout(p[, name])	Append a dropout layer to the block.
bidirectional()
forward(x)	Define the computation performed at every call.
get_default_normalization()	Returns the default normalization function for the block.
insert_dropout(p, after[, name])	Insert a dropout layer to the block.
prepend_dropout(p[, name])	Prepend a dropout layer to the block.
run_with_dummy_data()

Attributes Documentation

is_recurrent

Methods Documentation

GRU()

LSTM()

RNN()

append_dropout(p: float, name: str | None = 'dropout')

Append a dropout layer to the block.

Parameters

pfloat

The dropout probability.

nameOptional[str], optional

The name of the dropout layer, by default “dropout”.

bidirectional()

forward(x)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_default_normalization() → DeeplayModule

Returns the default normalization function for the block.

insert_dropout(p: float, after: str, name: str | None = 'dropout')

Insert a dropout layer to the block.

Parameters

pfloat

The dropout probability.

afterstr

The name of the layer after which the dropout layer will be executed.

nameOptional[str], optional

The name of the dropout layer, by default “dropout”.

Raises

ValueError

If the layer after is not found in the block.

prepend_dropout(p: float, name: str | None = 'dropout')

Prepend a dropout layer to the block.

Parameters

pfloat

The dropout probability.

nameOptional[str], optional

The name of the dropout layer, by default “dropout”.

run_with_dummy_data()