PIN-200M: A Knowledge-Intensive Dataset of Paired and Interleaved Multimodal Documents

PIN-200M contains approximately 200 million samples of paired and interleaved multimodal documents, requiring around 312 terabytes of storage. The dataset is a mini version of the PIN dataset introduced in a paper from June 2024. It was created by author m-a-p and last updated on Hugging Face in April 2026.

Use Cases

• Training multimodal large language models based on knowledge-intensive document pairs.
• Benchmarking model performance on interleaved text and image understanding tasks.
• Developing retrieval-augmented generation systems based on paired document structures.
• Conducting research on cross-modal alignment and reasoning using paired samples.

Strengths

• Contains approximately 200 million samples, indicating a large scale.
• Stores around 312 terabytes of data, suggesting high sample complexity or resolution.
• Includes quality signals for swift sample assessment, as mentioned in the description.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• License information is unavailable, which restricts clarity on permissible use.

Provenance

Source

huggingface

Collection Method

Method of gathering is not specified in the provided input.

Freshness

Last updated 2026-04-15 07:30:56; freshness should be verified.