Understanding Benchmarks -- Data Preparation

July 2, 2026 · View on GitHub

This guide covers data preparation for the understanding evaluation benchmarks: MME, MMMU, MMBench, MMStar, MM-Vet, and MathVista.

All benchmark data is stored under data/ at the repository root.

The data preparation follows the InternVL pipeline. For the full reference, see: https://internvl.readthedocs.io/en/latest/get_started/eval_data_preparation.html

MME

Download the MME benchmark data:

cd /path/to/umm_codebase
mkdir -p data/mme
cd data/mme
wget https://huggingface.co/OpenGVLab/InternVL/resolve/main/MME_Benchmark_release_version.zip
unzip MME_Benchmark_release_version.zip

Expected directory structure:

data/mme/
└── MME_Benchmark_release_version/
    ├── artwork/
    ├── celebrity/
    └── ...

Referenced in config as:

image_root: data/mme/MME_Benchmark_release_version

MMMU

MMMU is auto-downloaded from HuggingFace (MMMU/MMMU) during evaluation. No manual download is needed.

  • Cached at: data/MMMU/

Referenced in config as:

root: MMMU/MMMU
cache_dir: data/MMMU

MMBench

Download the MMBench TSV files:

cd /path/to/umm_codebase
mkdir -p data/mmbench
cd data/mmbench
wget https://download.openmmlab.com/mmclassification/datasets/mmbench/mmbench_dev_20230712.tsv
wget https://download.openmmlab.com/mmclassification/datasets/mmbench/mmbench_dev_cn_20231003.tsv
wget https://download.openmmlab.com/mmclassification/datasets/mmbench/mmbench_dev_en_20231003.tsv
wget https://download.openmmlab.com/mmclassification/datasets/mmbench/mmbench_test_cn_20231003.tsv
wget https://download.openmmlab.com/mmclassification/datasets/mmbench/mmbench_test_en_20231003.tsv

Expected directory structure:

data/mmbench/
├── mmbench_dev_20230712.tsv
├── mmbench_dev_cn_20231003.tsv
├── mmbench_dev_en_20231003.tsv
├── mmbench_test_cn_20231003.tsv
└── mmbench_test_en_20231003.tsv

MMStar

MMStar is auto-downloaded from HuggingFace (Lin-Chen/MMStar) during evaluation. No manual download is needed.

  • Cached at: data/MMStar/

Referenced in config as:

root: Lin-Chen/MMStar
split: val
cache_dir: data/MMStar

MM-Vet

Download the MM-Vet benchmark data and question file:

cd /path/to/umm_codebase
mkdir -p data/mm-vet
cd data/mm-vet
wget https://github.com/yuweihao/MM-Vet/releases/download/v1/mm-vet.zip
unzip mm-vet.zip
wget https://huggingface.co/OpenGVLab/InternVL/raw/main/llava-mm-vet.jsonl

Expected directory structure:

data/mm-vet/
├── mm-vet/
│   └── images/
│       ├── v1_0.png
│       └── ...
└── llava-mm-vet.jsonl

Referenced in config as:

image_root: data/mm-vet/mm-vet/images
question: data/mm-vet/llava-mm-vet.jsonl

MathVista

Download the annotation file:

cd /path/to/umm_codebase
mkdir -p data/MathVista
cd data/MathVista
wget https://huggingface.co/datasets/AI4Math/MathVista/raw/main/annot_testmini.json

Images are auto-downloaded from HuggingFace (AI4Math/MathVista) during evaluation.

Referenced in config as:

gt_file: data/MathVista/annot_testmini.json
cache_dir: data/MathVista