InsAVE80KVT2VRetrieval

An MTEB dataset

Massive Text Embedding Benchmark

Instruction-conditioned composed audio-video retrieval built from the held-out evaluation split of InsAVE-80K, the audio-video editing dataset released with InstructAV2AV. Each query pairs a source clip with the natural-language editing instruction that was applied to it, and the relevant document is the corresponding edited clip. The candidate pool is the complete set of evaluation clips, so every query's own source clip stays in the pool as a hard negative and the instruction is required to prefer the edited clip over the unedited one. Clips carry synchronised audio, and instructions may contain / markers delimiting intended spoken content. This is a new MOEB retrieval construction derived from the official 1,000-pair generation-evaluation split: InstructAV2AV scores generation quality on these pairs and defines no retrieval evaluation, so scores here are not comparable to any number reported in the paper.

Task category	Any2AnyRetrieval (video+text-to-video)
Domains	Scene, Spoken, Web
Reference	InstructAV2AV: Instruction-Guided Audio-Video Joint Editing

Source datasets:

• suimu/InsAVE-80K

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("InsAVE80KVT2VRetrieval")
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.

@article{zheng2026instructav2av,
  author = {Zheng, Haojie and Yang, Yixin and Yang, Siqi and Weng, Shuchen and Shi, Boxin},
  journal = {arXiv preprint arXiv:2605.18467},
  title = {InstructAV2AV: Instruction-Guided Audio-Video Joint Editing},
  year = {2026},
}


@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("InsAVE80KVT2VRetrieval")

desc_stats = task.metadata.descriptive_stats

{
    "test": {
        "num_samples": 3000,
        "num_queries": 1000,
        "num_documents": 2000,
        "number_of_characters": 156386,
        "documents_text_statistics": null,
        "documents_image_statistics": null,
        "documents_audio_statistics": null,
        "documents_video_statistics": {
            "total_duration_seconds": 9278.873660215679,
            "total_frames": 222954,
            "min_width": 1280,
            "average_width": 1280.0,
            "max_width": 1280,
            "min_height": 704,
            "average_height": 704.0,
            "max_height": 704,
            "min_duration_seconds": 2.7027027027027026,
            "average_duration_seconds": 4.639436830107839,
            "max_duration_seconds": 6.75,
            "unique_videos": 1872,
            "average_fps": 24.054,
            "fps": {
                "24": 1972,
                "25": 12,
                "30": 16
            },
            "min_resolution": [
                1280,
                704
            ],
            "average_resolution": [
                1280.0,
                704.0
            ],
            "max_resolution": [
                1280,
                704
            ],
            "resolutions": {
                "1280x704": 2000
            }
        },
        "queries_text_statistics": {
            "total_text_length": 156386,
            "min_text_length": 29,
            "average_text_length": 156.386,
            "max_text_length": 442,
            "unique_texts": 992
        },
        "queries_image_statistics": null,
        "queries_audio_statistics": null,
        "queries_video_statistics": {
            "total_duration_seconds": 4639.436830107839,
            "total_frames": 111477,
            "min_width": 1280,
            "average_width": 1280.0,
            "max_width": 1280,
            "min_height": 704,
            "average_height": 704.0,
            "max_height": 704,
            "min_duration_seconds": 2.7027027027027026,
            "average_duration_seconds": 4.639436830107839,
            "max_duration_seconds": 6.75,
            "unique_videos": 992,
            "average_fps": 24.054,
            "fps": {
                "24": 986,
                "25": 6,
                "30": 8
            },
            "min_resolution": [
                1280,
                704
            ],
            "average_resolution": [
                1280.0,
                704.0
            ],
            "max_resolution": [
                1280,
                704
            ],
            "resolutions": {
                "1280x704": 1000
            }
        },
        "relevant_docs_statistics": {
            "num_relevant_docs": 1120,
            "min_relevant_docs_per_query": 1,
            "average_relevant_docs_per_query": 1.12,
            "max_relevant_docs_per_query": 2,
            "unique_relevant_docs": 1120
        },
        "top_ranked_statistics": null
    }
}

---

This dataset card was automatically generated using MTEB