dataset

Datasets for Lightning IR that data loading and sampling.

This module defines several datasets that handle loading and sampling data for training and inference.

Classes

DocDataset(doc_dataset[, num_docs, text_fields])

IRDataset(dataset)

QueryDataset(query_dataset[, num_queries])

RunDataset(run_path_or_id[, depth, ...])

Sampler()

Helper class for sampling subsets of documents from a ranked list.

TupleDataset(tuples_dataset[, targets, num_docs])