nlp_architect.nn.torch.data package

Submodules

nlp_architect.nn.torch.data.dataset module

class nlp_architect.nn.torch.data.dataset.CombinedTensorDataset(datasets: List[torch.utils.data.dataset.TensorDataset])[source]

Bases: torch.utils.data.dataset.Dataset

Dataset as a concatenation of tensor datasets with different number of tensors (labeled dataset/ unlabeled dataset). Labels of unlabeled dataset will be represented as a tensor of zeros.

Each sample will be retrieved by indexing tensors along the first dimension.

Parameters:datasets (List[TensorDataset]) – datasets to concat.
class nlp_architect.nn.torch.data.dataset.ConcatTensorDataset(dataset: torch.utils.data.dataset.TensorDataset, datasets: List[torch.utils.data.dataset.TensorDataset])[source]

Bases: torch.utils.data.dataset.Dataset

Dataset as a concatenation of multiple TensorDataset datasets with same number of tensors.

Each sample will be retrieved by indexing tensors along the first dimension.

Parameters:
  • dataset (TensorDataset) – dataset to which rest datasets will be concatinated.
  • datasets (List[TensorDataset]) – datasets to concat to the dataset.
class nlp_architect.nn.torch.data.dataset.ParallelDataset(*datasets)[source]

Bases: torch.utils.data.dataset.Dataset

Module contents