MBench-Data / README.md
studyOverflow's picture
Publish canonical MBench-A and MBench-T dataset release
412581e
|
Raw
History Blame Contribute Delete
983 Bytes
---
pretty_name: MBench Data
---
# MBench Data
This repository contains the evaluation data released with
[MBench](https://github.com/study-overflow/MBench).
- `MBench-A`: 547 action-conditioned evaluation samples.
- `MBench-T`: 493 base cases, expanded into 684 condition-level evaluation items.
## Structure
Each benchmark follows the directory protocol expected by the MBench evaluator:
```text
MBench-{A|T}/
├── dataset.yaml
├── samples/{subset}/{sample_id}/sample.json
└── models/{model_id}/
├── model.yaml
├── samples.jsonl
└── outputs/{subset}/{sample_id}/{condition_id}/video.mp4
```
The exact evaluation items available for each model are listed in
`models/{model_id}/samples.jsonl`.
In MBench-T, the 493 base cases consist of 293 non-causal samples and 200
causal samples. The causal samples are expanded into 200 camera-motion, 100
occlusion, and 91 dimming conditions, producing 684 condition-level items in
total.