Dataset
patents_claims_1.5m_traim_test
About 1.5 million US patent claims split into training and test partitions in CSV, organized for claim-level classification and summarization work (Apache 2.0). A large, ready-split corpus for IP-claims modeling.
Data preview
Dataset card
The Patents Claims 1.5M dataset, published by AI-Growth-Lab, is a tabular collection of roughly 1.5 million patent claim records distributed as a CSV. It pairs free-text claims with per-row indicators for Cooperative Patent Classification codes, and the repository's metadata is published with no stated license.
What this dataset contains
Each row corresponds to a single patent claim, identified by an integer id field and accompanied by a date string and the claim text. The remaining columns are integer flags for IPC/CPC subclasses, beginning with agriculture and food-related codes such as A01B through A01L (soil working, planting, harvesting, and animal husbandry) and continuing into food and beverage subclasses including A21B through A23V. The repository reports 1,492,294 total rows split between a train and a test partition.
Structure and fields
The schema lists an int64 id, a string date, a string text field holding the claim, and a long tail of int64 columns named after classification subclasses, with A01B, A01C, A01D, A01F, A01G, A01H, A01J, A01K and A01L explicitly enumerated alongside A21B, A21C, A21D, A22B, A22C, A23B, A23C, A23D, A23F, A23G, A23J, A23K, A23L, A23N, A23P and A23V. Tags indicate a CSV format with tabular and text modalities, a size band of 1M to 10M rows, and loading support for pandas, polars and mlcroissant.
How analysts use it
Because the rows pair a claim text with subclass flags such as A01B, A21C or A23L, the dataset is suited to training and benchmarking multi-label classification of patent claims against classification codes, or to text-mining claim language by technological area. Investment analysts could in principle use the per-class counts to monitor activity in agriculture (A01B–A01N) or food processing (A21B–A23V) as a public-records proxy for innovation trends, though the absence of a stated license and explicit row counts by subclass limits its suitability for production use.
Documentation & references 1
Related datasets
Browse moreParsed eligibility criteria from 13,229 ClinicalTrials.gov studies representing the candidate pool gathered by two first-stage retrievers during TREC Clinical Trials 2021–2023, formatted as typed entity-relation graphs for reranking models.
The clinical-trials-patient-graphs dataset is a small tabular collection of patient-level clinical records distributed in Parquet format by the 2001jdev publisher. It organizes entities and relations extracted from trial data into structured rows spanning 2021 through 2023.
This dataset is a parsed snapshot of ClinicalTrials.gov records, distributed in Parquet format on the Hugging Face Hub under an unspecified license. It contains roughly 52,000 trial entries drawn from public registry filings.
clinical-trials-trec-qrels is a tabular relevance-judgment file distributed by the 2001jdev user on the Hugging Face Hub. It maps clinical-trial topic identifiers to NCT registry IDs with graded relevance scores.
The clinical-trials-trec-topics dataset is a small tabular collection of clinical-trial topic records distributed in Parquet format on the Hugging Face Hub. It appears to be derived from TREC clinical-trial retrieval topics, with rows partitioned by year.
SEC EDGAR 10-K & 10-Q financial statements for 2,600+ NASDAQ companies.
Discussion
Questions or observations about this dataset? Join the conversation on X.
No comments yet. Start the conversation on X.