MINT-1T: A Multimodal Dataset with One Trillion Text Tokens and 3.4 Billion Images

MINT-1T is an open-source multimodal interleaved dataset designed for pretraining research. It contains one trillion text tokens and 3.4 billion images, representing a 10x scale-up from prior open-source collections and includes sources like PDFs and arXiv papers. The dataset was created by a team from the University of Washington and was last updated on the platform in September 2024.

Use Cases

• Multimodal model pretraining based on the interleaved text and image data.
• Scaling up training data for vision-language models based on the one trillion token and 3.4 billion image scale.
• Research on incorporating academic and document sources based on the inclusion of PDFs and arXiv papers.

Strengths

• Contains one trillion text tokens and 3.4 billion images, a 10x scale-up from existing open-source datasets.
• Includes previously untapped sources such as PDFs and arXiv papers.
• Explicitly designed for multimodal pretraining research.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and exact license details are unknown, which may limit suitability assessment.

Provenance

Source

University of Washington team (mlfoundations)

Collection Method

Aggregated from multiple open data sources, including PDFs and arXiv papers.

Time Range

Likely includes content up to 2024, but specific temporal coverage is not stated.

Freshness

Last updated 2024-09-19 21:02:55; freshness should be verified.

Geography

Region tag indicates 'us', but specific spatial coverage is not detailed.