datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
neomme-eval-vidore-v3-industrial
NeoMME Evaluation Sample — ViDoRe v3 Industrial
A 760-page, 75-query evaluation sample drawn from
vidore/vidore_v3_industrial
(revision e26c864), built for benchmarking
NeoMME-260M-Retriever
and other visual document retrieval models.
Contents
Split
Items
Notes
data/pages/
760 PNG images
Scanned USAF technical manual pages (1000x1600 RGB)
data/corpus.json
760 rows
corpus_id, doc_id, page_number_in_doc
data/queries.jsonl
75 queries
Natural language… See the full description on the dataset page: https://huggingface.co/datasets/b4ph/neomme-eval-vidore-v3-industrial.vidore3_finance_neomme_260m_li
vidore3_finance_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe finance (vidore/finance), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_finance_en at revision 7f432c176d82e27546501ad8064a713ac3071809, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids,
unchanged.
Every document is… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_finance_neomme_260m_li.vidore3_financefr_neomme_260m_li
vidore3_financefr_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe financefr (vidore/financefr), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_finance_fr at revision 1d808daa08032ffecdf62da151a7f7a8fe2bd0c9, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids,
unchanged.
Every… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_financefr_neomme_260m_li.vidore3_hr_neomme_260m_li
vidore3_hr_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe hr (vidore/hr), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_hr at revision 0cdf0979f2c5a0fd3e335e6373b9da48a9fe3bc3, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids,
unchanged.
Every document is one variable-length… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_hr_neomme_260m_li.vidore3_computerscience_neomme_260m_li
vidore3_computerscience_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe computerscience (vidore/computerscience), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_computer_science at revision d5cc75883d92e294f0c0fc2662551c9708a06ebc, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_computerscience_neomme_260m_li.vidore3_pharmaceuticals_neomme_260m_li
vidore3_pharmaceuticals_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe pharmaceuticals (vidore/pharmaceuticals), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_pharmaceuticals at revision 3abd4aa8a9445fb5538a78a19ba50bd57bd22b5c, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_pharmaceuticals_neomme_260m_li.vidore3_physics_neomme_260m_li
vidore3_physics_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe physics (vidore/physics), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_physics at revision a0de276f515acc044b72cae8de53a44bb5a8f1f5, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids,
unchanged.
Every document is… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_physics_neomme_260m_li.vidore3_energy_neomme_260m_li
vidore3_energy_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe energy (vidore/energy), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_energy at revision caec06d3c73434d635f710f93bcd898331c59f20, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids,
unchanged.
Every document is one… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_energy_neomme_260m_li.vidore3_industrial_neomme_260m_li
vidore3_industrial_neomme_260m_li
Multi-vector (late-interaction) embeddings of ViDoRe industrial (vidore/industrial), encoded with
Hcompany/NeoMME-260M-Retriever-ST-late at revision 023be2a8ab9d797f5aa76f5bf8b5dde78d819659.
Source data: Hugging Face dataset vidore/vidore_v3_industrial at revision e26c864724f5dd71a3d7d739272d95637764cee9, configs corpus / queries / qrels, split test, loaded with datasets. Document, query and qrel ids are the source's own ids,
unchanged.
Every… See the full description on the dataset page: https://huggingface.co/datasets/robro612/vidore3_industrial_neomme_260m_li.