neuralmonkey.encoders.numpy_stateful_filler module

class neuralmonkey.encoders.numpy_stateful_filler.SpatialFiller(name: str, input_shape: List[int], data_id: str, projection_dim: int = None, ff_hidden_dim: int = None, reuse: neuralmonkey.model.model_part.ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: List[Tuple[str, Callable]] = None) → None

Bases: neuralmonkey.model.model_part.ModelPart, neuralmonkey.model.stateful.SpatialStatefulWithOutput

Placeholder class for 3D numerical input.

This model part is used to feed 3D tensors (e.g., pre-trained convolutional maps image captioning). Optionally, the states are projected to given size.

__init__(name: str, input_shape: List[int], data_id: str, projection_dim: int = None, ff_hidden_dim: int = None, reuse: neuralmonkey.model.model_part.ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: List[Tuple[str, Callable]] = None) → None

Instantiate SpatialFiller.

Parameters:
  • name – Name of the model part.
  • input_shape – Dimensionality of the input.
  • data_id – Name of the data series with numpy objects.
  • projection_dim – Optional, dimension of the states projection.
feed_dict(dataset: neuralmonkey.dataset.Dataset, train: bool = False) → Dict[tensorflow.python.framework.ops.Tensor, Any]

Return a feed dictionary for the given feedable object.

Parameters:
  • dataset – A dataset instance from which to get the data.
  • train – Boolean indicating whether the model runs in training mode.
Returns:

A FeedDict dictionary object.
output

Return the object output.

A 2D Tensor of shape (batch, state_size) which contains the resulting state of the object.

spatial_mask

Return mask for the spatial_states.

A 3D Tensor of shape (batch, width, height) of type float32 which masks the spatial states that they can be of different shapes. The mask should only contain ones or zeros.

spatial_states

Return object states in space.

A 4D Tensor of shape (batch, width, height, state_size) which contains the states of the object in space (e.g. final layer of a convolution network processing an image.

class neuralmonkey.encoders.numpy_stateful_filler.StatefulFiller(name: str, dimension: int, data_id: str, output_shape: int = None, reuse: neuralmonkey.model.model_part.ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: List[Tuple[str, Callable]] = None) → None

Bases: neuralmonkey.model.model_part.ModelPart, neuralmonkey.model.stateful.Stateful

Placeholder class for stateful input.

This model part is used to feed 1D tensors to the model. Optionally, it projects the states to given dimension.

__init__(name: str, dimension: int, data_id: str, output_shape: int = None, reuse: neuralmonkey.model.model_part.ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: List[Tuple[str, Callable]] = None) → None

Instantiate StatefulFiller.

Parameters:
  • name – Name of the model part.
  • dimension – Dimensionality of the input.
  • data_id – Series containing the numpy objects.
  • output_shape – Dimension of optional state projection.
feed_dict(dataset: neuralmonkey.dataset.Dataset, train: bool = False) → Dict[tensorflow.python.framework.ops.Tensor, Any]

Return a feed dictionary for the given feedable object.

Parameters:
  • dataset – A dataset instance from which to get the data.
  • train – Boolean indicating whether the model runs in training mode.
Returns:

A FeedDict dictionary object.
output

Return the object output.

A 2D Tensor of shape (batch, state_size) which contains the resulting state of the object.