DSSMTrainDataset

class rectools.dataset.torch_datasets.DSSMTrainDataset(items: csr_matrix, users: csr_matrix, interactions: csr_matrix)[source]

Bases: DSSMTrainDatasetBase

Torch dataset wrapper for rectools.dataset.dataset.Dataset. Implements torch.utils.data.Dataset for subsequent usage with torch.utils.data.DataLoader. Does the following: for a given index takes a row of user interactions, a row of user features and samples one positive and one negative items and then returns them as tensors.

This class is intended for internal usage or advanced users who want to implement more sophisticated sampling logic.

Parameters

  • items (csr_matrix) – Item features.

  • users (csr_matrix) – User features.

  • interactions (csr_matrix) – Interactions matrix.

Inherited-members

Methods

from_dataset(dataset)