MMTEB: Massive Multilingual Text Embedding Benchmark
Paper • 2502.13595 • Published • 50
Datasets:
id string | text string |
|---|---|
0 | a video of brush hair |
1 | a video of cartwheel |
2 | a video of catch |
3 | a video of chew |
4 | a video of clap |
5 | a video of climb |
6 | a video of climb stairs |
7 | a video of dive |
8 | a video of draw sword |
9 | a video of dribble |
10 | a video of drink |
11 | a video of eat |
12 | a video of fall floor |
13 | a video of fencing |
14 | a video of flic flac |
15 | a video of golf |
16 | a video of handstand |
17 | a video of hit |
18 | a video of hug |
19 | a video of jump |
20 | a video of kick |
21 | a video of kick ball |
22 | a video of kiss |
23 | a video of laugh |
24 | a video of pick |
25 | a video of pour |
26 | a video of pullup |
27 | a video of punch |
28 | a video of push |
29 | a video of pushup |
30 | a video of ride bike |
31 | a video of ride horse |
32 | a video of run |
33 | a video of shake hands |
34 | a video of shoot ball |
35 | a video of shoot bow |
36 | a video of shoot gun |
37 | a video of sit |
38 | a video of situp |
39 | a video of smile |
40 | a video of smoke |
41 | a video of somersault |
42 | a video of stand |
43 | a video of swing baseball |
44 | a video of sword |
45 | a video of sword exercise |
46 | a video of talk |
47 | a video of throw |
48 | a video of turn |
49 | a video of walk |
50 | a video of wave |
HMDB51 is a large video database for human motion recognition with 51 action categories from digitized movies and online sources. Used official split 1 across 51 action classes (~3,570 train / ~1,530 test).
| Task category | VideoZeroshotClassification (video-to-text) |
| Domains | Activity, Web |
| Reference | 2011 International Conference on Computer Vision |
Source datasets:
How to evaluate on this task
You can evaluate an embedding model on this dataset using the following code:
import mteb
task = mteb.get_task("HMDB51ZeroShot")
model = mteb.get_model(YOUR_MODEL)
mteb.evaluate(model, task)
To learn more about how to run models on mteb task check out the GitHub repository.
Citation
If you use this dataset, please cite the dataset as well as mteb, as this dataset likely includes additional processing as a part of the MMTEB Contribution.
@inproceedings{kuehne2011hmdb,
author = {Kuehne, Hildegard and Jhuang, Hueihan and Garrote, Est{\'\i}baliz and Poggio, Tomaso and Serre, Thomas},
booktitle = {2011 International Conference on Computer Vision},
organization = {IEEE},
pages = {2556--2563},
title = {HMDB: a large video database for human motion recognition},
year = {2011},
}
@article{enevoldsen2025mmtebmassivemultilingualtext,
title={MMTEB: Massive Multilingual Text Embedding Benchmark},
author={Kenneth Enevoldsen and Isaac Chung and Imene Kerboua and Márton Kardos and Ashwin Mathur and David Stap and Jay Gala and Wissam Siblini and Dominik Krzemiński and Genta Indra Winata and Saba Sturua and Saiteja Utpala and Mathieu Ciancone and Marion Schaeffer and Gabriel Sequeira and Diganta Misra and Shreeya Dhakal and Jonathan Rystrøm and Roman Solomatin and Ömer Çağatan and Akash Kundu and Martin Bernstorff and Shitao Xiao and Akshita Sukhlecha and Bhavish Pahwa and Rafał Poświata and Kranthi Kiran GV and Shawon Ashraf and Daniel Auras and Björn Plüster and Jan Philipp Harries and Loïc Magne and Isabelle Mohr and Mariya Hendriksen and Dawei Zhu and Hippolyte Gisserot-Boukhlef and Tom Aarsen and Jan Kostkan and Konrad Wojtasik and Taemin Lee and Marek Šuppa and Crystina Zhang and Roberta Rocca and Mohammed Hamdy and Andrianos Michail and John Yang and Manuel Faysse and Aleksei Vatolin and Nandan Thakur and Manan Dey and Dipam Vasani and Pranjal Chitale and Simone Tedeschi and Nguyen Tai and Artem Snegirev and Michael Günther and Mengzhou Xia and Weijia Shi and Xing Han Lù and Jordan Clive and Gayatri Krishnakumar and Anna Maksimova and Silvan Wehrli and Maria Tikhonova and Henil Panchal and Aleksandr Abramov and Malte Ostendorff and Zheng Liu and Simon Clematide and Lester James Miranda and Alena Fenogenova and Guangyu Song and Ruqiya Bin Safi and Wen-Ding Li and Alessia Borghini and Federico Cassano and Hongjin Su and Jimmy Lin and Howard Yen and Lasse Hansen and Sara Hooker and Chenghao Xiao and Vaibhav Adlakha and Orion Weller and Siva Reddy and Niklas Muennighoff},
publisher = {arXiv},
journal={arXiv preprint arXiv:2502.13595},
year={2025},
url={https://arxiv.org/abs/2502.13595},
doi = {10.48550/arXiv.2502.13595},
}
@article{muennighoff2022mteb,
author = {Muennighoff, Niklas and Tazi, Nouamane and Magne, Loïc and Reimers, Nils},
title = {MTEB: Massive Text Embedding Benchmark},
publisher = {arXiv},
journal={arXiv preprint arXiv:2210.07316},
year = {2022}
url = {https://arxiv.org/abs/2210.07316},
doi = {10.48550/ARXIV.2210.07316},
}
Dataset Statistics
Dataset Statistics
The following code contains the descriptive statistics from the task. These can also be obtained using:
import mteb
task = mteb.get_task("HMDB51ZeroShot")
desc_stats = task.metadata.descriptive_stats
{
"test": {
"num_samples": 1530,
"text_statistics": null,
"image_statistics": null,
"audio_statistics": null,
"video_statistics": {
"total_duration_seconds": 4833.7666850000005,
"total_frames": 145013,
"min_width": 288,
"average_width": 359.01437908496735,
"max_width": 592,
"min_height": 240,
"average_height": 240.0,
"max_height": 240,
"min_duration_seconds": 0.733333,
"average_duration_seconds": 3.1593246307189546,
"max_duration_seconds": 21.6,
"unique_videos": 1530,
"average_fps": 30.0,
"fps": {
"30": 1530
},
"min_resolution": [
288,
240
],
"average_resolution": [
359.01437908496735,
240.0
],
"max_resolution": [
592,
240
],
"resolutions": {
"320x240": 885,
"352x240": 158,
"416x240": 103,
"424x240": 42,
"480x240": 29,
"360x240": 32,
"288x240": 29,
"304x240": 4,
"560x240": 79,
"440x240": 3,
"428x240": 7,
"432x240": 62,
"448x240": 6,
"576x240": 12,
"592x240": 14,
"300x240": 26,
"348x240": 6,
"312x240": 14,
"332x240": 2,
"316x240": 1,
"292x240": 3,
"420x240": 7,
"568x240": 1,
"308x240": 2,
"452x240": 1,
"496x240": 1,
"400x240": 1
}
},
"label_statistics": {
"min_labels_per_text": 1,
"average_label_per_text": 1.0,
"max_labels_per_text": 1,
"unique_labels": 51,
"labels": {
"0": {
"count": 30
},
"1": {
"count": 30
},
"2": {
"count": 30
},
"3": {
"count": 30
},
"4": {
"count": 30
},
"5": {
"count": 30
},
"6": {
"count": 30
},
"7": {
"count": 30
},
"8": {
"count": 30
},
"9": {
"count": 30
},
"10": {
"count": 30
},
"11": {
"count": 30
},
"12": {
"count": 30
},
"13": {
"count": 30
},
"14": {
"count": 30
},
"15": {
"count": 30
},
"16": {
"count": 30
},
"17": {
"count": 30
},
"18": {
"count": 30
},
"19": {
"count": 30
},
"20": {
"count": 30
},
"21": {
"count": 30
},
"22": {
"count": 30
},
"23": {
"count": 30
},
"24": {
"count": 30
},
"25": {
"count": 30
},
"26": {
"count": 30
},
"27": {
"count": 30
},
"28": {
"count": 30
},
"29": {
"count": 30
},
"30": {
"count": 30
},
"31": {
"count": 30
},
"32": {
"count": 30
},
"33": {
"count": 30
},
"34": {
"count": 30
},
"35": {
"count": 30
},
"36": {
"count": 30
},
"37": {
"count": 30
},
"38": {
"count": 30
},
"39": {
"count": 30
},
"40": {
"count": 30
},
"41": {
"count": 30
},
"42": {
"count": 30
},
"43": {
"count": 30
},
"44": {
"count": 30
},
"45": {
"count": 30
},
"46": {
"count": 30
},
"47": {
"count": 30
},
"48": {
"count": 30
},
"49": {
"count": 30
},
"50": {
"count": 30
}
}
},
"candidates_labels_text_statistics": {
"total_text_length": 887,
"min_text_length": 14,
"average_text_length": 17.392156862745097,
"max_text_length": 25,
"unique_texts": 51
}
}
}
This dataset card was automatically generated using MTEB
- Downloads last month
- 130