Patents Claims: 1.5 Million Records for Training and Testing

A dataset of patent claims, likely containing 1.5 million records intended for machine learning training and testing. It was published on the Hugging Face platform by the user aipaa1512 and was last updated on April 30, 2026. The specific source and collection method for the patent documents are not detailed in the provided metadata.

Use Cases

• Training a language model to generate or classify patent claims (inferred from domain, verify after download)
• Fine-tuning a text similarity model for prior art search (inferred from domain, verify after download)
• Developing a sequence labeling model to extract key entities from patent text (inferred from domain, verify after download)

Strengths

• Published on the Hugging Face platform.
• Last updated on 2026-04-30 21:05:01.

Limitations

• Metadata is minimal; actual content requires verification after download.
• Column names, data types, and sample data are unknown, making pre-download assessment difficult.
• The license, organization of origin, and exact data collection method are unspecified.

Provenance

Source

huggingface

Collection Method

unknown

Time Range

unknown

Freshness

Last updated 2026-04 30 21:05:01.

Geography

unknown