{"url":"/dataset/ms-marco","name":"MS MARCO","full_name":"Microsoft Machine Reading Comprehension Dataset","description_markdown":"The **MS MARCO** (Microsoft MAchine Reading Comprehension) is a collection of datasets focused on deep learning in search.\r\nThe first dataset was a question answering dataset featuring 100,000 real Bing questions and a human generated answer. Over time the collection was extended with a 1,000,000 question dataset, a natural language generation dataset, a passage ranking dataset, keyphrase extraction dataset, crawling dataset, and a conversational search.\r\n\r\nSource: [https://microsoft.github.io/msmarco/](https://microsoft.github.io/msmarco/)\r\nImage Source: [https://arxiv.org/pdf/1809.08267.pdf](https://arxiv.org/pdf/1809.08267.pdf)","description_withheld":null,"homepage":"https://microsoft.github.io/msmarco/","introduced_date":"2016-01-01","introduced_date_note":null,"introduced_by":{"paper":"/paper/ms-marco-a-human-generated-machine-reading","title":"MS MARCO: A Human Generated MAchine Reading COmprehension Dataset","first_author":"Payal Bajaj","url":null},"license":{"name":"Custom (research-only, non-commercial)","url":"https://microsoft.github.io/msmarco/#:~:text=Terms%20and%20Conditions"},"modalities":[{"name":"Texts","url":"/datasets/modality/texts"}],"tasks":[{"name":"Question Answering","url":"/task/question-answering","datasets_with_task":"/datasets/task/question-answering"},{"name":"Information Retrieval","url":"/task/information-retrieval","datasets_with_task":"/datasets/task/information-retrieval"},{"name":"Reading Comprehension","url":"/task/reading-comprehension","datasets_with_task":"/datasets/task/reading-comprehension"},{"name":"Passage Retrieval","url":"/task/passage-retrieval","datasets_with_task":"/datasets/task/passage-retrieval"},{"name":"Text Retrieval","url":"/task/text-retrieval","datasets_with_task":"/datasets/task/text-retrieval"},{"name":"Passage Re-Ranking","url":"/task/passage-re-ranking","datasets_with_task":"/datasets/task/passage-re-ranking"},{"name":"TREC 2019 Passage Ranking","url":"/task/trec-2019-passage-ranking","datasets_with_task":"/datasets/task/trec-2019-passage-ranking"},{"name":"Passage Ranking","url":"/task/passage-ranking","datasets_with_task":"/datasets/task/passage-ranking"}],"languages":[{"name":"English","url":"/datasets/language/english"}],"variants":["MS MARCO","MSMARCO"],"data_loaders":[{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/microsoft/ms_marco","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/ms_marco","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/facebookresearch/ParlAI","url":"https://parl.ai/docs/tasks.html#ms_marco","frameworks":["pytorch"]}],"num_papers_in_archive":1036,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/passage-ranking-on-ms-marco","task":"Passage Ranking","dataset_variant":"MS MARCO","rows":4,"metrics":["MRR@10"],"first_row_in_archive_order":{"model":"Fine-tuned SOTA","paper":"/paper/text-and-code-embeddings-by-contrastive-pre","metrics":{"MRR@10":"44.3"},"code_links":[{"title":"openmatch/coco-dr","url":"https://github.com/openmatch/coco-dr"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/passage-re-ranking-on-ms-marco","task":"Passage Re-Ranking","dataset_variant":"MS MARCO","rows":4,"metrics":["MRR"],"first_row_in_archive_order":{"model":"HLATR","paper":"/paper/hlatr-enhance-multi-stage-text-retrieval-with","metrics":{"MRR":"0.42"},"code_links":[{"title":"Alibaba-NLP/HLATR","url":"https://github.com/Alibaba-NLP/HLATR"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/question-answering-on-ms-marco","task":"Question Answering","dataset_variant":"MS MARCO","rows":4,"metrics":["Rouge-L","BLEU-1"],"first_row_in_archive_order":{"model":"Masque Q&A Style","paper":"/paper/multi-style-generative-reading-comprehension","metrics":{"BLEU-1":"43.77","Rouge-L":"52.2"},"code_links":[]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/information-retrieval-on-ms-marco","task":"Information Retrieval","dataset_variant":"MS MARCO","rows":3,"metrics":["Time (ms)","MRR@10"],"first_row_in_archive_order":{"model":"ConAE-128","paper":"/paper/dimension-reduction-for-efficient-dense","metrics":{"Time (ms)":"0.3245"},"code_links":[{"title":"neuir/conae","url":"https://github.com/neuir/conae"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/information-retrieval-on-msmarco","task":"Information Retrieval","dataset_variant":"MSMARCO","rows":1,"metrics":["MRR@10"],"first_row_in_archive_order":{"model":"RetroMAE","paper":"/paper/retromae-pre-training-retrieval-oriented","metrics":{"MRR@10":"0.416"},"code_links":[{"title":"staoxiao/retromae","url":"https://github.com/staoxiao/retromae"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/passage-retrieval-on-ms-marco-1","task":"Passage Retrieval","dataset_variant":"MS MARCO","rows":1,"metrics":["MAP"],"first_row_in_archive_order":{"model":"TW-BERT","paper":"/paper/end-to-end-query-term-weighting","metrics":{"MAP":"0.2025"},"code_links":[{"title":"ledmaster/tw-bert","url":"https://github.com/ledmaster/tw-bert"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/text-retrieval-on-ms-marco","task":"Text Retrieval","dataset_variant":"MS MARCO","rows":1,"metrics":["NDCG@10"],"first_row_in_archive_order":{"model":"Lucene (BM25S)","paper":"/paper/bm25s-orders-of-magnitude-faster-lexical","metrics":{"NDCG@10":"22.8"},"code_links":[{"title":"xhluca/bm25s","url":"https://github.com/xhluca/bm25s"},{"title":"xhluca/bm25-benchmarks","url":"https://github.com/xhluca/bm25-benchmarks"},{"title":"conda-forge/bm25s-feedstock","url":"https://github.com/conda-forge/bm25s-feedstock"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/bm25s-orders-of-magnitude-faster-lexical","title":"BM25S: Orders of magnitude faster lexical search via eager sparse scoring","date":"2024-07-04","rows_on_this_dataset":1,"code_links":3,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":22,"samples_ran":10,"samples_unverified":12,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/end-to-end-query-term-weighting","title":"End-to-End Query Term Weighting","date":"2023-08-06","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/retromae-v2-duplex-masked-auto-encoder-for","title":"RetroMAE v2: Duplex Masked Auto-Encoder For Pre-Training Retrieval-Oriented Language Models","date":"2022-11-16","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/retromae-pre-training-retrieval-oriented","title":"RetroMAE: Pre-Training Retrieval-oriented Language Models Via Masked Auto-Encoder","date":"2022-05-24","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/hlatr-enhance-multi-stage-text-retrieval-with","title":"HLATR: Enhance Multi-stage Text Retrieval with Hybrid List Aware Transformer Reranking","date":"2022-05-21","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/dimension-reduction-for-efficient-dense","title":"Dimension Reduction for Efficient Dense Retrieval via Conditional Autoencoder","date":"2022-05-06","rows_on_this_dataset":2,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":1,"samples_ran":1,"samples_unverified":0,"pointer_only_for_licence":1,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/text-and-code-embeddings-by-contrastive-pre","title":"Text and Code Embeddings by Contrastive Pre-Training","date":"2022-01-24","rows_on_this_dataset":4,"code_links":1,"syntology":null},{"paper":"/paper/document-expansion-by-query-prediction","title":"Document Expansion by Query Prediction","date":"2019-04-17","rows_on_this_dataset":1,"code_links":5,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":3,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/an-updated-duet-model-for-passage-re-ranking","title":"An Updated Duet Model for Passage Re-ranking","date":"2019-03-18","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/passage-re-ranking-with-bert","title":"Passage Re-ranking with BERT","date":"2019-01-13","rows_on_this_dataset":1,"code_links":6,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":13,"samples_ran":2,"samples_unverified":11,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/multi-style-generative-reading-comprehension","title":"Multi-style Generative Reading Comprehension","date":"2019-01-08","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/a-deep-cascade-model-for-multi-document","title":"A Deep Cascade Model for Multi-Document Reading Comprehension","date":"2018-11-28","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/multi-passage-machine-reading-comprehension","title":"Multi-Passage Machine Reading Comprehension with Cross-Passage Answer Verification","date":"2018-05-06","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/bidirectional-attention-flow-for-machine","title":"Bidirectional Attention Flow for Machine Comprehension","date":"2016-11-05","rows_on_this_dataset":1,"code_links":27,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":11,"samples_ran":8,"samples_unverified":3,"pointer_only_for_licence":7,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":5,"samples_harvested":50,"samples_ran":24,"samples_unverified":26,"pointer_only_for_licence":8,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}