ddro
Datasets
All datasets matching “ddro”ddro-nq-dataset
DDRO — NQ320K Processed Dataset
This dataset contains the preprocessed Natural Questions (NQ320K) corpus used to train and evaluate the DDRO generative retrieval models from:
📄 Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval (SIGIR 2025)
The raw NQ data (from Google) is processed into a unified format with document text, queries, and relevance annotations, ready for use in generative retrieval training pipelines.
Files… See the full description on the dataset page: https://huggingface.co/datasets/kiyam/ddro-nq-dataset.ddro-docids
📄 ddro-docids
This repository provides the generated document IDs (DocIDs) used for training and evaluating the DDRO (Direct Document Relevance Optimization) models.
Two types of DocIDs are included:
PQ (Product Quantization) DocIDs: Compact semantic representations based on quantized document embeddings.
TU (Title + URL) DocIDs: Tokenized document identifiers constructed from document titles and/or URLs.
📚 Contents
pq_msmarco_docids.txt: PQ DocIDs for MS… See the full description on the dataset page: https://huggingface.co/datasets/kiyam/ddro-docids.ddro-msmarco-doc-dataset-300k
DDRO — MS MARCO Top-300K Processed Dataset
This dataset contains the preprocessed MS MARCO Top-300K document corpus used to train and evaluate the DDRO generative retrieval models from:
Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval (SIGIR 2025)
Files
File
Description
Size
msmarco-docs-sents.top.300k.json
Top-300K documents selected by click frequency, with sentence tokenization (JSONL format)
~2 GB… See the full description on the dataset page: https://huggingface.co/datasets/kiyam/ddro-msmarco-doc-dataset-300k.ddro-testsetsddro-pseudo-queries
📄 ddro-pseudo-queries
This repository contains the generated pseudo queries used for training the DDRO (Direct Document Relevance Optimization) models.The pseudo queries were created following the DocTTTTTQuery approach to expand the training data for generative document retrieval.
📚 Contents
pseudo_queries_msmarco.txt: Pseudo queries for the MS MARCO (MS300K) dataset.
pseudo_queries_nq.txt: Pseudo queries for the Natural Questions (NQ320K) dataset.
Each file maps… See the full description on the dataset page: https://huggingface.co/datasets/kiyam/ddro-pseudo-queries.nq_ddro_reproduction_with_genir_datasets