maskit.datasets package¶
Submodules¶
maskit.datasets.circles module¶
maskit.datasets.iris module¶
maskit.datasets.mnist module¶
- maskit.datasets.mnist.apply_PCA(wires: int, x_train: ndarray)¶
- maskit.datasets.mnist.convert_label(y: int, classes: List[int]) List[float] ¶
- maskit.datasets.mnist.downscale(x_data, y_data, size) Tuple[ndarray, ndarray] ¶
This function does several things including the reduction of image size, conversion to binary as well as removal of duplicates.
- Parameters:
x_data – Input data
y_data – Target data
size – Maximum number of data to prepare
- maskit.datasets.mnist.mnist(wires: int = 4, classes=(6, 9), train_size: int = 100, validation_size: int = 0, test_size: int = 50, shuffle: Union[bool, int] = 1337) DataSet ¶
- maskit.datasets.mnist.prepare_data(x_data: ndarray, y_data: ndarray, size, classes) Tuple[ndarray, ndarray] ¶
- maskit.datasets.mnist.reduce_image(x: ndarray) ndarray ¶
maskit.datasets.utils module¶
- class maskit.datasets.utils.DataSet(train_data, train_target, validation_data, validation_target, test_data, test_target)¶
Bases:
tuple
- property test_data¶
Alias for field number 4
- property test_target¶
Alias for field number 5
- property train_data¶
Alias for field number 0
- property train_target¶
Alias for field number 1
- property validation_data¶
Alias for field number 2
- property validation_target¶
Alias for field number 3
- maskit.datasets.utils.one_hot(a: ndarray, num_classes: int) ndarray ¶
- maskit.datasets.utils.pad_data(data: ndarray, axis: int, padding: int) ndarray ¶
Function pads 0 data to the end of the given axis.
- Parameters:
data – The data to pad 0 data to
axis – Axis to pad on
padding – How many elements to pad
Module contents¶
- maskit.datasets.load_data(dataset: str, train_size: int = 100, test_size: int = 50, validation_size: int = 0, shuffle: Union[bool, int] = 1337, classes: Tuple[int, ...] = (6, 9), wires: int = 4, target_length: Optional[int] = None) DataSet ¶
Returns the data for the requested
dataset
.- Parameters:
dataset – Name of the chosen dataset. Available datasets are: iris, mnist and circles.
train_size – Size of the training dataset
test_size – Size of the testing dataset
shuffle – if the dataset should be shuffled, used also as a seed
classes – which numbers of the mnist dataset should be included
wires – number of wires in the circuit
target_length – Normalised length of target arrays
- Raises:
ValueError – Raised if a not supported dataset is requested