Datasets:
Tasks:
Question Answering
Modalities:
Audio
Formats:
soundfolder
Languages:
English
Size:
1K - 10K
ArXiv:
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -13,9 +13,6 @@ MMSU, a comprehensive benchmark designed specifically for understanding and reas
|
|
| 13 |
MMSU comprises 5,000 meticulously curated audio-question-answer triplets across 47 distinct tasks.
|
| 14 |
To ground our benchmark in linguistic theory, we systematically incorporate a wide range of linguistic phenomena, including phonetics, prosody, syntax, syntactics, semantics, and paralinguistics.
|
| 15 |
|
| 16 |
-
## Evaluation:
|
| 17 |
-
[**Refer to Github Code**](https://github.com/dingdongwang/MMSU_Bench/tree/main)
|
| 18 |
-
|
| 19 |
## Citation
|
| 20 |
```
|
| 21 |
@article{wang2025mmsu,
|
|
|
|
| 13 |
MMSU comprises 5,000 meticulously curated audio-question-answer triplets across 47 distinct tasks.
|
| 14 |
To ground our benchmark in linguistic theory, we systematically incorporate a wide range of linguistic phenomena, including phonetics, prosody, syntax, syntactics, semantics, and paralinguistics.
|
| 15 |
|
|
|
|
|
|
|
|
|
|
| 16 |
## Citation
|
| 17 |
```
|
| 18 |
@article{wang2025mmsu,
|