← Back to API reference

neuralbench.main.Data

pydantic model neuralbench.main.Data

Create dataloaders for brain-modeling experiments.

Fields:

• batch_size (int)

• channel_positions (neuralset.extractors.neuro.ChannelPositions | None)

• drop_incomplete (bool)

• drop_last (bool)

• duration (float | None)

• min_finite_target_fraction (float | None)

• neuro (neuralset.extractors.base.BaseExtractor)

• num_workers (int)

• persistent_workers (bool)

• pin_memory (bool)

• prefetch_factor (int | None)

• sampler (neuralbench.data.BaseSampler | None)

• seed (int | None)

• start (float)

• stream_by (list[str] | None)

• stride (float | None)

• stride_drop_incomplete (bool)

• study (neuralset.base.Step)

• summary_columns (list[str])

• target (neuralset.extractors.base.BaseExtractor)

• test_batch_size (int | None)

• test_shuffle (bool)

• train_shuffle (bool)

• trigger_event_type (str | list[str])

• val_batch_size (int | None)

• val_shuffle (bool)

field study: Step [Required]

field neuro: BaseExtractor [Required]

field target: BaseExtractor [Required]

field channel_positions: ChannelPositions | None = None

field trigger_event_type: str | list[str] [Required]

field start: float = -0.5

field duration: float | None = 3

field stride: float | None = None

field stride_drop_incomplete: bool = True

field drop_incomplete: bool = False

field min_finite_target_fraction: float | None = None

field sampler: BaseSampler | None = None

field batch_size: int = 64

field val_batch_size: int | None = None

field test_batch_size: int | None = None

field train_shuffle: bool = True

field val_shuffle: bool = False

field test_shuffle: bool = False

field num_workers: int = 0

field drop_last: bool = False

field pin_memory: bool = True

field persistent_workers: bool = True

field prefetch_factor: int | None = None

field seed: int | None = None

field stream_by: list[str] | None = None

field summary_columns: list[str] = []

prepare() → dict[str, DataLoader]

Load events, build extractors, segment data and return train/val/test DataLoaders.

Returns:

• dict with keys "train", "val", "test" mapping to

• DataLoader instances.

← Back to API reference