qualia_plugin_snn.dataset package

Submodules

• qualia_plugin_snn.dataset.DVSGesture module
  • DVSGesture
    • DVSGesture.h
    • DVSGesture.w
    • DVSGesture.dtype
    • DVSGesture.__init__()
    • DVSGesture.__call__()
    • DVSGesture.name
• qualia_plugin_snn.dataset.DVSGestureWithPreprocessing module
  • DVSGestureWithPreprocessing
    • DVSGestureWithPreprocessing.__init__()
    • DVSGestureWithPreprocessing.__call__()
    • DVSGestureWithPreprocessing.name
• qualia_plugin_snn.dataset.EventDataset module
  • EventDataset
    • EventDataset.h
    • EventDataset.w
    • EventDataset.__init__()
    • EventDataset.import_data()
  • EventDatasetChunks
    • EventDatasetChunks.h
    • EventDatasetChunks.w
    • EventDatasetChunks.__init__()
    • EventDatasetChunks.import_data()
• qualia_plugin_snn.dataset.SHD module
  • SHD
    • SHD.h
    • SHD.w
    • SHD.__init__()
    • SHD.__call__()
• qualia_plugin_snn.dataset.SSC module
  • SSC
    • SSC.__init__()
    • SSC.__call__()

Module contents

Qualia-Plugin-SNN dataset package contains event-based or spiking datasets.

class qualia_plugin_snn.dataset.SHD

Bases: EventDataset

Spiking Heidelberg Digits dataset data loading.

h: int = 1

w: int = 700

__init__(path: str = '', prefix: str = 'shd') → None

Instantiate the Spiking Heidelberg Digits dataset loader.

Parameters:

• path (str) – Dataset source path

• prefix (str) – source file name prefix, default to shd

Return type:

None

__call__() → EventDataModel

Load Spiking Heidelberge Digits data as events.

Returns:

Data model structure with train and test sets containing events and labels

Return type:

EventDataModel

class qualia_plugin_snn.dataset.SSC

Bases: SHD

Spiking Speech Commands dataset data loading.

__init__(path: str = '', prefix: str = 'ssc') → None

Instantiate the Spiking Speech Commands dataset loader.

Parameters:

• path (str) – Dataset source path

• prefix (str) – source file name prefix, default to ssc

Return type:

None

__call__() → EventDataModel

Load Spiking Heidelberge Digits data as events.

Returns:

Data model structure with train and test sets containing events and labels

Return type:

EventDataModel

class qualia_plugin_snn.dataset.DVSGesture

Bases: EventDatasetChunks

DVS128 Gesture event-based data loading based on SpikingJelly.

h: int = 128

w: int = 128

dtype: Final[dtype] = dtype([('t', '<i8'), ('y', 'i1'), ('x', 'i1'), ('p', '?')])

__init__(path: str = '', data_type: str = 'frame') → None

Instantiate the DVS128 Gesture dataset loader.

Parameters:

• path (str) – Dataset source path

• data_type (str) – Only 'frame' is supported

Return type:

None

__call__() → EventDataChunksModel

Load DVS128 Gesture data as events.

Returns:

Data model structure with train and test sets containing events and labels

Return type:

EventDataChunksModel

property name: str

class qualia_plugin_snn.dataset.DVSGestureWithPreprocessing

Bases: RawDataset

DVS128 Gesture event-based data loading based on SpikingJelly including preprocessing to frames and timesteps.

__init__(path: str = '', data_type: str = 'frame', duration: int = 0, timesteps: int = 0) → None

Instantiate the DVS128 Gesture dataset loader with preprocessing.

Parameters:

• path (str) – Dataset source path

• data_type (str) – Only 'frame' is supported

• duration (int) – Frame integration duration

• timesteps (int) – Number of timesteps to groupe frames by

Return type:

None

__call__() → RawDataModel

Load DVS128 Gesture data as frames over timesteps.

Returns:

Data model structure with train and test sets containing frames with timesteps and labels

Return type:

RawDataModel

property name: str