search

MMMU-Benchmark / MMMU

This repo contains evaluation code for the paper "MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI"
https://mmmu-benchmark.github.io/
Apache License 2.0
327 stars 21 forks source link

Enquiry about the usage about your dataset #33

Closed tunantu closed 1 week ago

tunantu commented 2 weeks ago

Hi, authors, thanks for your work. I am curious about the usage about your dataset, i.e., how to load the dataset, how to get the corresponding text and image. Meanwhile, I had looked throught the page about your hugging face, unfortunately, I did not get it. So could you release a simple example about how to use this dataset?

Many thanks.

xiangyue9607 commented 2 weeks ago

Thank you! The script is now available at https://github.com/MMMU-Benchmark/MMMU/tree/main/mmmu-pro