Massive Text Embedding Benchmark

Installation | Usage | Leaderboard | Documentation | Citing

Installation

pip install mteb

Usage

• Using a python script (see scripts/run_mteb_english.py and mteb/mtebscripts for more):

import mteb
from sentence_transformers import SentenceTransformer

# Define the sentence-transformers model name
model_name = "average_word_embeddings_komninos"
# or directly from huggingface:
# model_name = "sentence-transformers/all-MiniLM-L6-v2"

model = SentenceTransformer(model_name)
tasks = mteb.get_tasks(tasks=["Banking77Classification"])
evaluation = mteb.MTEB(tasks=tasks)
results = evaluation.run(model, output_folder=f"results/{model_name}")

• Using CLI

mteb available_tasks

mteb run -m sentence-transformers/all-MiniLM-L6-v2 \
    -t Banking77Classification  \
    --verbosity 3

# if nothing is specified default to saving the results in the results/{model_name} folder

• Using multiple GPUs in parallel can be done by just having a custom encode function that distributes the inputs to multiple GPUs like e.g. here or here.

Advanced Usage

Click on each section below to see the details.

Dataset selection

Dataset selection

Datasets can be selected by providing the list of datasets, but also

• by their task (e.g. "Clustering" or "Classification")

tasks = mteb.get_tasks(task_types=["Clustering", "Retrieval"]) # Only select clustering and retrieval tasks

• by their categories e.g. "s2s" (sentence to sentence) or "p2p" (paragraph to paragraph)

tasks = mteb.get_tasks(categories=["s2s", "p2p"]) # Only select sentence2sentence and paragraph2paragraph datasets

• by their languages

tasks = mteb.get_tasks(languages=["eng", "deu"]) # Only select datasets which contain "eng" or "deu" (iso 639-3 codes)

You can also specify which languages to load for multilingual/cross-lingual tasks like below:

import mteb

tasks = [
    mteb.get_task("AmazonReviewsClassification", languages = ["eng", "fra"]),
    mteb.get_task("BUCCBitextMining", languages = ["deu"]), # all subsets containing "deu"
]

# or you can select specific huggingface subsets like this:
from mteb.tasks import AmazonReviewsClassification, BUCCBitextMining

evaluation = mteb.MTEB(tasks=[
        AmazonReviewsClassification(hf_subsets=["en", "fr"]) # Only load "en" and "fr" subsets of Amazon Reviews
        BUCCBitextMining(hf_subsets=["de-en"]), # Only load "de-en" subset of BUCC
])
# for an example of a HF subset see "Subset" in the dataset viewer at: https://huggingface.co/datasets/mteb/bucc-bitext-mining

There are also presets available for certain task collections, e.g. to select the 56 English datasets that form the "Overall MTEB English leaderboard":

from mteb import MTEB_MAIN_EN
evaluation = mteb.MTEB(tasks=MTEB_MAIN_EN, task_langs=["en"])

Passing in `encode` arguments

Passing in encode arguments

To pass in arguments to the model's encode function, you can use the encode keyword arguments (encode_kwargs):

evaluation.run(model, encode_kwargs={"batch_size": 32}

Selecting evaluation split

Selecting evaluation split

You can evaluate only on test splits of all tasks by doing the following:

evaluation.run(model, eval_splits=["test"])

Note that the public leaderboard uses the test splits for all datasets except MSMARCO, where the "dev" split is used.

Using a custom model

Using a custom model

Models should implement the following interface, implementing an encode function taking as inputs a list of sentences, and returning a list of embeddings (embeddings can be np.array, torch.tensor, etc.). For inspiration, you can look at the mteb/mtebscripts repo used for running diverse models via SLURM scripts for the paper.

class MyModel():
    def encode(
        self, sentences: list[str], **kwargs: Any
    ) -> torch.Tensor | np.ndarray:
        """Encodes the given sentences using the encoder.

        Args:
            sentences: The sentences to encode.
            **kwargs: Additional arguments to pass to the encoder.

        Returns:
            The encoded sentences.
        """
        pass

model = MyModel()
tasks = mteb.get_task("Banking77Classification")
evaluation = MTEB(tasks=tasks)
evaluation.run(model)

If you'd like to use different encoding functions for query and corpus when evaluating on Retrieval or Reranking tasks, you can add separate methods for encode_queries and encode_corpus. If these methods exist, they will be automatically used for those tasks. You can refer to the DRESModel at mteb/evaluation/evaluators/RetrievalEvaluator.py for an example of these functions.

class MyModel():
    def encode_queries(self, queries: list[str], **kwargs) -> list[np.ndarray] | list[torch.Tensor]:
        """
        Returns a list of embeddings for the given sentences.
        Args:
            queries: List of sentences to encode

        Returns:
            List of embeddings for the given sentences
        """
        pass

    def encode_corpus(self, corpus: list[str] | list[dict[str, str]], **kwargs) -> list[np.ndarray] | list[torch.Tensor]:
        """
        Returns a list of embeddings for the given sentences.
        Args:
            corpus: List of sentences to encode
                or list of dictionaries with keys "title" and "text"

        Returns:
            List of embeddings for the given sentences
        """
        pass

Evaluating on a custom dataset

Evaluating on a custom dataset

To evaluate on a custom task, you can run the following code on your custom task. See how to add a new task, for how to create a new task in MTEB.

from mteb import MTEB
from mteb.abstasks.AbsTaskReranking import AbsTaskReranking
from sentence_transformers import SentenceTransformer


class MyCustomTask(AbsTaskReranking):
    ...

model = SentenceTransformer("average_word_embeddings_komninos")
evaluation = MTEB(tasks=[MyCustomTask()])
evaluation.run(model)

Using a cross encoder for reranking

Using a cross encoder for reranking

To use a cross encoder for reranking, you can directly use a CrossEncoder from SentenceTransformers. The following code shows a two-stage run with the second stage reading results saved from the first stage.

from mteb import MTEB
import mteb
from sentence_transformers import CrossEncoder, SentenceTransformer

cross_encoder = CrossEncoder("cross-encoder/ms-marco-TinyBERT-L-2-v2")
dual_encoder = SentenceTransformer("all-MiniLM-L6-v2")

tasks = mteb.get_tasks(tasks=["NFCorpus"], languages=["eng"])

subset = "default" # subset name used in the NFCorpus dataset
eval_splits = ["test"]

evaluation = MTEB(tasks=tasks)
evaluation.run(
    dual_encoder,
    eval_splits=eval_splits,
    save_predictions=True,
    output_folder="results/stage1",
)
evaluation.run(
    cross_encoder,
    eval_splits=eval_splits,
    top_k=5,
    save_predictions=True,
    output_folder="results/stage2",
    previous_results=f"results/stage1/NFCorpus_{subset}_predictions.json",
)

Saving retrieval task predictions

Saving retrieval task predictions

To save the predictions from a retrieval task, add the --save_predictions flag in the CLI or set save_predictions=True in the run method. The filename will be in the "{task_name}_{subset}_predictions.json" format.

Python:

from mteb import MTEB
import mteb
from sentence_transformers import SentenceTransformer

model = SentenceTransformer("all-MiniLM-L6-v2")

tasks = mteb.get_tasks( tasks=["NFCorpus"], languages=["eng"])

evaluation = MTEB(tasks=tasks)
evaluation.run(
    model,
    eval_splits=["test"],
    save_predictions=True,
    output_folder="results",
)

CLI:

mteb run -t NFCorpus -m all-MiniLM-L6-v2 --output_folder results --save_predictions

Fetching result from the results repository

Multiple models have already been run on tasks avaiable within MTEB. These results are available results repository.

To make the results more easily accecible we have designed custom functionality for retrieving from the repository. For instance, you are selecting the best model for your French and English retrieval task on legal documents you could fetch the relevant tasks and create a dataframe of the results using the following code:

import mteb
from mteb.task_selection import results_to_dataframe

tasks = mteb.get_tasks(
    task_types=["Retrieval"], languages=["eng", "fra"], domains=["Legal"]
)

model_names = [
    "GritLM/GritLM-7B",
    "intfloat/multilingual-e5-small",
    "intfloat/multilingual-e5-base",
    "intfloat/multilingual-e5-large",
]
models = [mteb.get_model_meta(name) for name in model_names]

results = mteb.load_results(models=models, tasks=tasks)

df = results_to_dataframe(results)

Documentation

Documentation
📋 Tasks	Overview of available tasks
📈 Leaderboard	The interactive leaderboard of the benchmark
🤖 Adding a model	Information related to how to submit a model to the leaderboard
👩‍🔬 Reproducible workflows	Information related to how to reproduce and create reproducible workflows with MTEB
👩‍💻 Adding a dataset	How to add a new task/dataset to MTEB
👩‍💻 Adding a leaderboard tab	How to add a new leaderboard tab to MTEB
🤝 Contributing	How to contribute to MTEB and set it up for development
🌐 MMTEB	An open-source effort to extend MTEB to cover a broad set of languages

Citing

MTEB was introduced in "MTEB: Massive Text Embedding Benchmark", feel free to cite:

@article{muennighoff2022mteb,
  doi = {10.48550/ARXIV.2210.07316},
  url = {https://arxiv.org/abs/2210.07316},
  author = {Muennighoff, Niklas and Tazi, Nouamane and Magne, Lo{\"\i}c and Reimers, Nils},
  title = {MTEB: Massive Text Embedding Benchmark},
  publisher = {arXiv},
  journal={arXiv preprint arXiv:2210.07316},  
  year = {2022}
}

You may also want to read and cite the amazing work that has extended MTEB & integrated new datasets:

• Shitao Xiao, Zheng Liu, Peitian Zhang, Niklas Muennighoff. "C-Pack: Packaged Resources To Advance General Chinese Embedding" arXiv 2023
• Michael Günther, Jackmin Ong, Isabelle Mohr, Alaeddine Abdessalem, Tanguy Abel, Mohammad Kalim Akram, Susana Guzman, Georgios Mastrapas, Saba Sturua, Bo Wang, Maximilian Werk, Nan Wang, Han Xiao. "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents" arXiv 2023
• Silvan Wehrli, Bert Arnrich, Christopher Irrgang. "German Text Embedding Clustering Benchmark" arXiv 2024
• Orion Weller, Benjamin Chang, Sean MacAvaney, Kyle Lo, Arman Cohan, Benjamin Van Durme, Dawn Lawrie, Luca Soldaini. "FollowIR: Evaluating and Teaching Information Retrieval Models to Follow Instructions" arXiv 2024
• Dawei Zhu, Liang Wang, Nan Yang, Yifan Song, Wenhao Wu, Furu Wei, Sujian Li. "LongEmbed: Extending Embedding Models for Long Context Retrieval" arXiv 2024
• Kenneth Enevoldsen, Márton Kardos, Niklas Muennighoff, Kristoffer Laigaard Nielbo. "The Scandinavian Embedding Benchmarks: Comprehensive Assessment of Multilingual and Monolingual Text Embedding" arXiv 2024

For works that have used MTEB for benchmarking, you can find them on the leaderboard.