iapp/MMMU-Thai

MMMU Thai (MMMU Benchmark Translated to Thai)

MMMU Thai is a dataset for evaluating multimodal models on massive multi-discipline tasks requiring college-level knowledge and deliberate reasoning. This dataset is translated from MMMU (A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI) into Thai.

Dataset Details

MMMU Thai consists of 11,500 meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines:

1. Art & Design
2. Business
3. Science
4. Health & Medicine
5. Humanities & Social Science
6. Tech & Engineering

These questions span 30 subjects and 183 subfields, comprising 30 highly heterogeneous image types, such as charts, diagrams, maps, tables, music sheets, and chemical structures.

Dataset Structure

MMMU Thai includes:

• Development set: 150 samples
• Validation set: 900 samples
• Test set: 10,500 questions (without answers)

The development set is used for few-shot/in-context learning, and the validation set is used for debugging models, selecting hyperparameters, or quick evaluations. The answers and explanations for the test set questions are withheld.

How We Built This Dataset

This dataset was automatically translated to Thai using openthaigpt1.5-72b on the columns: question, options, and explanation, ensuring consistency between options and answers using human. Improved using Qwen.

LICENSE

This dataset is dual-licensed under Apache License 2.0 and the Qwen LICENSE Agreement. The original MMMU dataset is licensed under Apache 2.0, while the improvements made using Qwen-derived models are subject to the Qwen LICENSE Agreement.

Maintainer

Kobkrit Viriyayudhakorn (kobkrit@iapp.co.th)

References

Original dataset: MMMU Dataset