pytorch_pfn_extras.training.extensions.variable_statistics_plot.Reservoir

class pytorch_pfn_extras.training.extensions.variable_statistics_plot.Reservoir(size, data_shape, dtype=<class 'numpy.float32'>)

Bases: object

Reservoir sample with a fixed sized buffer.

Methods

__init__(size, data_shape[, dtype])
add(x[, idx])
get_data()

Parameters:

• size (int) –

• data_shape (Tuple[int, ...]) –

• dtype (Any) –

__init__(size, data_shape, dtype=<class 'numpy.float32'>)

Parameters:

• size (int) –

• data_shape (Tuple[int, ...]) –

• dtype (Any) –

Return type:

None

add(x, idx=None)

Parameters:

• x (Any) –

• idx (Optional[Any]) –

Return type:

None

get_data()

Return type:

Tuple[Any, Any]