Large-Scale Multimodal Chemical Structure Images

MolParse v1.0 is a multimodal dataset released in January 2026 for optical chemical structure parsing. It contains a large-scale collection of molecular structure images sourced from scientific literature, designed to train models that convert diagrams into structured chemical representations.

Use Cases

• Train multimodal models to parse molecular structure images into structured chemical representations like SMILES or InChI.
• Benchmark optical chemical recognition (OCR) performance on realistic chemical diagrams from scientific literature.
• Develop image-to-text models specialized for converting complex chemical diagrams into machine-readable formats.

Strengths

• Large-scale collection of molecular structure images.
• Focuses on realistic chemical diagrams from scientific literature.
• Released in January 2026, indicating recent data curation.

Limitations

• Specific scale metrics like row count, image count, or column details are unknown.
• Dataset scope is limited to chemical structure parsing, not general molecular property prediction.

Provenance

Source

InnovatorLab on Hugging Face.

Collection Method

Images of chemical diagrams sourced from scientific literature.

Time Range

null

Freshness

Last updated 2026-02-06.

Geography

null