"""MTEB Results""" from __future__ import annotations import json import datasets logger = datasets.logging.get_logger(__name__) _CITATION = """@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} } """ _DESCRIPTION = """Results on MTEB""" URL = "https://huggingface.co/datasets/mteb/results/resolve/main/paths.json" VERSION = datasets.Version("1.0.1") EVAL_LANGS = [ "af", "afr-eng", "am", "amh", "amh-eng", "ang-eng", "ar", "ar-ar", "ara-eng", "arq-eng", "arz-eng", "ast-eng", "awa-eng", "az", "aze-eng", "bel-eng", "ben-eng", "ber-eng", "bn", "bos-eng", "bre-eng", "bul-eng", "cat-eng", "cbk-eng", "ceb-eng", "ces-eng", "cha-eng", "cmn-eng", "cor-eng", "csb-eng", "cy", "cym-eng", "da", "dan-eng", "de", "de-fr", "de-pl", "deu-eng", "dsb-eng", "dtp-eng", "el", "ell-eng", "en", "en-ar", "en-de", "en-en", "en-tr", "eng", "epo-eng", "es", "es-en", "es-es", "es-it", "est-eng", "eus-eng", "fa", "fao-eng", "fi", "fin-eng", "fr", "fr-en", "fr-pl", "fra", "fra-eng", "fry-eng", "gla-eng", "gle-eng", "glg-eng", "gsw-eng", "hau", "he", "heb-eng", "hi", "hin-eng", "hrv-eng", "hsb-eng", "hu", "hun-eng", "hy", "hye-eng", "ibo", "id", "ido-eng", "ile-eng", "ina-eng", "ind-eng", "is", "isl-eng", "it", "it-en", "ita-eng", "ja", "jav-eng", "jpn-eng", "jv", "ka", "kab-eng", "kat-eng", "kaz-eng", "khm-eng", "km", "kn", "ko", "ko-ko", "kor-eng", "kur-eng", "kzj-eng", "lat-eng", "lfn-eng", "lit-eng", "lin", "lug", "lv", "lvs-eng", "mal-eng", "mar-eng", "max-eng", "mhr-eng", "mkd-eng", "ml", "mn", "mon-eng", "ms", "my", "nb", "nds-eng", "nl", "nl-ende-en", "nld-eng", "nno-eng", "nob-eng", "nov-eng", "oci-eng", "orm", "orv-eng", "pam-eng", "pcm", "pes-eng", "pl", "pl-en", "pms-eng", "pol-eng", "por-eng", "pt", "ro", "ron-eng", "ru", "run", "rus-eng", "sl", "slk-eng", "slv-eng", "spa-eng", "sna", "som", "sq", "sqi-eng", "srp-eng", "sv", "sw", "swa", "swe-eng", "swg-eng", "swh-eng", "ta", "tam-eng", "tat-eng", "te", "tel-eng", "tgl-eng", "th", "tha-eng", "tir", "tl", "tr", "tuk-eng", "tur-eng", "tzl-eng", "uig-eng", "ukr-eng", "ur", "urd-eng", "uzb-eng", "vi", "vie-eng", "war-eng", "wuu-eng", "xho", "xho-eng", "yid-eng", "yor", "yue-eng", "zh", "zh-CN", "zh-TW", "zh-en", "zsm-eng", ] # v_measures key is somehow present in voyage-2-law results and is a list SKIP_KEYS = ["std", "evaluation_time", "main_score", "threshold", "v_measures", "scores_per_experiment"] # Use "train" split instead TRAIN_SPLIT = ["DanishPoliticalCommentsClassification"] # Use "validation" split instead VALIDATION_SPLIT = [ "AFQMC", "Cmnli", "IFlyTek", "LEMBSummScreenFDRetrieval", "MSMARCO", "MSMARCO-PL", "MultilingualSentiment", "Ocnli", "TNews", ] # Use "dev" split instead DEV_SPLIT = [ "CmedqaRetrieval", "CovidRetrieval", "DuRetrieval", "EcomRetrieval", "MedicalRetrieval", "MMarcoReranking", "MMarcoRetrieval", "MSMARCO", "MSMARCO-PL", "T2Reranking", "T2Retrieval", "VideoRetrieval", "TERRa", "MIRACLReranking", "MIRACLRetrieval", ] # Use "test.full" split TESTFULL_SPLIT = ["OpusparcusPC"] # Use "standard" split STANDARD_SPLIT = ["BrightRetrieval"] # Use "devtest" split DEVTEST_SPLIT = ["FloresBitextMining"] TEST_AVG_SPLIT = { "LEMBNeedleRetrieval": [ "test_256", "test_512", "test_1024", "test_2048", "test_4096", "test_8192", "test_16384", "test_32768", ], "LEMBPasskeyRetrieval": [ "test_256", "test_512", "test_1024", "test_2048", "test_4096", "test_8192", "test_16384", "test_32768", ], } MODELS = [ "Alibaba-NLP__gte-Qwen1.5-7B-instruct", "Alibaba-NLP__gte-Qwen2-7B-instruct", "BAAI__bge-base-en", "BAAI__bge-base-en-v1.5", "BAAI__bge-base-en-v1.5-instruct", "BAAI__bge-base-zh", "BAAI__bge-base-zh-v1.5", "BAAI__bge-large-en", "BAAI__bge-large-en-v1.5", "BAAI__bge-large-en-v1.5-instruct", "BAAI__bge-large-zh", "BAAI__bge-large-zh-noinstruct", "BAAI__bge-large-zh-v1.5", "BAAI__bge-m3", "BAAI__bge-m3-instruct", "BAAI__bge-small-en-v1.5", "BAAI__bge-small-en-v1.5-instruct", "BAAI__bge-small-zh", "BAAI__bge-small-zh-v1.5", "Cohere__Cohere-embed-english-v3.0", "Cohere__Cohere-embed-english-v3.0-instruct", "Cohere__Cohere-embed-multilingual-light-v3.0", "Cohere__Cohere-embed-multilingual-v3.0", "DeepPavlov__distilrubert-small-cased-conversational", "DeepPavlov__rubert-base-cased", "DeepPavlov__rubert-base-cased-sentence", "FacebookAI__xlm-roberta-base", "FacebookAI__xlm-roberta-large", "Geotrend__bert-base-10lang-cased", "Geotrend__bert-base-15lang-cased", "Geotrend__bert-base-25lang-cased", "Geotrend__distilbert-base-25lang-cased", "Geotrend__distilbert-base-en-fr-cased", "Geotrend__distilbert-base-en-fr-es-pt-it-cased", "Geotrend__distilbert-base-fr-cased", "GritLM__GritLM-7B", "GritLM__GritLM-7B-noinstruct", "KBLab__electra-small-swedish-cased-discriminator", "KBLab__sentence-bert-swedish-cased", "KB__bert-base-swedish-cased", "McGill-NLP__LLM2Vec-Llama-2-7b-chat-hf-mntp-supervised", "McGill-NLP__LLM2Vec-Llama-2-unsupervised", "McGill-NLP__LLM2Vec-Meta-Llama-3-supervised", "McGill-NLP__LLM2Vec-Meta-Llama-3-unsupervised", "McGill-NLP__LLM2Vec-Mistral-supervised", "McGill-NLP__LLM2Vec-Mistral-unsupervised", "McGill-NLP__LLM2Vec-Sheared-Llama-supervised", "McGill-NLP__LLM2Vec-Sheared-Llama-unsupervised", "Muennighoff__SGPT-1.3B-weightedmean-msmarco-specb-bitfit", "Muennighoff__SGPT-125M-weightedmean-msmarco-specb-bitfit", "Muennighoff__SGPT-125M-weightedmean-msmarco-specb-bitfit-doc", "Muennighoff__SGPT-125M-weightedmean-msmarco-specb-bitfit-que", "Muennighoff__SGPT-125M-weightedmean-nli-bitfit", "Muennighoff__SGPT-2.7B-weightedmean-msmarco-specb-bitfit", "Muennighoff__SGPT-5.8B-weightedmean-msmarco-specb-bitfit", "Muennighoff__SGPT-5.8B-weightedmean-msmarco-specb-bitfit-que", "Muennighoff__SGPT-5.8B-weightedmean-nli-bitfit", "NbAiLab__nb-bert-base", "NbAiLab__nb-bert-large", "Salesforce__SFR-Embedding-Mistral", "T-Systems-onsite__cross-en-de-roberta-sentence-transformer", "Wissam42__sentence-croissant-llm-base", "ai-forever__sbert_large_mt_nlu_ru", "ai-forever__sbert_large_nlu_ru", "aliyun__OpenSearch-text-hybrid", "almanach__camembert-base", "almanach__camembert-large", "amazon__titan-embed-text-v1", "baichuan-ai__text-embedding", "bigscience-data__sgpt-bloom-1b7-nli", "bigscience-data__sgpt-bloom-7b1-msmarco", "bm25", "bm25s", "castorini__monobert-large-msmarco", "castorini__monot5-3b-msmarco-10k", "castorini__monot5-base-msmarco-10k", "chcaa__dfm-encoder-large-v1", "cointegrated__LaBSE-en-ru", "cointegrated__rubert-tiny", "cointegrated__rubert-tiny2", "dangvantuan__sentence-camembert-base", "dangvantuan__sentence-camembert-large", "deepfile__embedder-100p", "deepset__gbert-base", "deepset__gbert-large", "deepset__gelectra-base", "deepset__gelectra-large", "deepvk__USER-base", "deepvk__USER-bge-m3", "deepvk__deberta-v1-base", "distilbert__distilbert-base-uncased", "dwzhu__e5-base-4k", "elastic__elser-v2", "facebook__contriever", "facebook__contriever-instruct", "facebook__dpr-ctx_encoder-multiset-base", "facebook__dragon-plus-context-encoder", "facebook__tart-full-flan-t5-xl", "facebookresearch__LASER2", "facebookresearch__dragon-plus", "facebookresearch__dragon-plus-instruct", "flaubert__flaubert_base_cased", "flaubert__flaubert_base_uncased", "flaubert__flaubert_large_cased", "google-bert__bert-base-multilingual-cased", "google-bert__bert-base-multilingual-uncased", "google-bert__bert-base-uncased", "google-gecko__text-embedding-004", "google-gecko__text-embedding-004-256", "google__flan-t5-base", "google__flan-t5-large", "hkunlp__instructor-base", "hkunlp__instructor-large", "hkunlp__instructor-xl", "intfloat__e5-base", "intfloat__e5-base-v2", "intfloat__e5-large", "intfloat__e5-large-v2", "intfloat__e5-mistral-7b-instruct", "intfloat__e5-mistral-7b-instruct-noinstruct", "intfloat__e5-small", "intfloat__e5-small-v2", "intfloat__multilingual-e5-base", "intfloat__multilingual-e5-large", "intfloat__multilingual-e5-large-instruct", "intfloat__multilingual-e5-small", "ipipan__herbert-base-retrieval-v2", "ipipan__silver-retriever-base-v1", "izhx__udever-bloom-1b1", "izhx__udever-bloom-560m", "jhu-clsp__FollowIR-7B", "jinaai__jina-embeddings-v2-base-en", "jonfd__electra-small-nordic", "ltg__norbert3-base", "ltg__norbert3-large", "meta-llama__llama-2-7b-chat", "mistral__mistral-embed", "mistralai__mistral-7b-instruct-v0.2", "mixedbread-ai__mxbai-embed-large-v1", "moka-ai__m3e-base", "moka-ai__m3e-large", "nomic-ai__nomic-embed-text-v1", "nomic-ai__nomic-embed-text-v1.5-128", "nomic-ai__nomic-embed-text-v1.5-256", "nomic-ai__nomic-embed-text-v1.5-512", "nomic-ai__nomic-embed-text-v1.5-64", "nthakur__contriever-base-msmarco", "openai__text-embedding-3-large", "openai__text-embedding-3-large-256", "openai__text-embedding-3-large-instruct", "openai__text-embedding-3-small-instruct", "openai__text-embedding-ada-002", "openai__text-embedding-ada-002-instruct", "openai__text-search-ada-001", "openai__text-search-ada-doc-001", "openai__text-search-babbage-001", "openai__text-search-curie-001", "openai__text-search-davinci-001", "openai__text-similarity-ada-001", "openai__text-similarity-babbage-001", "openai__text-similarity-curie-001", "openai__text-similarity-davinci-001", "openai__text-embedding-3-small", "orionweller__tart-dual-contriever-msmarco", "princeton-nlp__sup-simcse-bert-base-uncased", "princeton-nlp__unsup-simcse-bert-base-uncased", "sdadas__st-polish-paraphrase-from-distilroberta", "sdadas__st-polish-paraphrase-from-mpnet", "sentence-transformers__LaBSE", "sentence-transformers__all-MiniLM-L12-v2", "sentence-transformers__all-MiniLM-L6-v2", "sentence-transformers__all-MiniLM-L6-v2-instruct", "sentence-transformers__all-mpnet-base-v2", "sentence-transformers__all-mpnet-base-v2-instruct", "sentence-transformers__allenai-specter", "sentence-transformers__average_word_embeddings_glove.6B.300d", "sentence-transformers__average_word_embeddings_komninos", "sentence-transformers__distiluse-base-multilingual-cased-v2", "sentence-transformers__gtr-t5-base", "sentence-transformers__gtr-t5-large", "sentence-transformers__gtr-t5-xl", "sentence-transformers__gtr-t5-xxl", "sentence-transformers__msmarco-bert-co-condensor", "sentence-transformers__multi-qa-MiniLM-L6-cos-v1", "sentence-transformers__paraphrase-multilingual-MiniLM-L12-v2", "sentence-transformers__paraphrase-multilingual-mpnet-base-v2", "sentence-transformers__sentence-t5-base", "sentence-transformers__sentence-t5-large", "sentence-transformers__sentence-t5-xl", "sentence-transformers__sentence-t5-xxl", "sentence-transformers__use-cmlm-multilingual", "sergeyzh__LaBSE-ru-turbo", "sergeyzh__rubert-tiny-turbo", "shibing624__text2vec-base-chinese", "shibing624__text2vec-base-multilingual", "shibing624__text2vec-large-chinese", "silk-road__luotuo-bert-medium", "uklfr__gottbert-base", "vesteinn__DanskBERT", "voyageai__voyage-2", "voyageai__voyage-code-2", "voyageai__voyage-large-2-instruct", "voyageai__voyage-law-2", "voyageai__voyage-lite-01-instruct", "voyageai__voyage-lite-02-instruct", "voyageai__voyage-multilingual-2", "vprelovac__universal-sentence-encoder-multilingual-3", "vprelovac__universal-sentence-encoder-multilingual-large-3", ] def get_model_for_current_dir(dir_name: str) -> str | None: for model in MODELS: if model == dir_name or ("__" in dir_name and dir_name.split("__")[1] == model): return model return None # Needs to be run whenever new files are added def get_paths(): import collections, json, os files = collections.defaultdict(list) for model_dir in os.listdir("results"): results_model_dir = os.path.join("results", model_dir) if not os.path.isdir(results_model_dir): print(f"Skipping {results_model_dir}") continue model_name = get_model_for_current_dir(model_dir) if model_name is None: print(f"Skipping {model_dir} model dir") continue for revision_folder in os.listdir(results_model_dir): if not os.path.isdir(os.path.join(results_model_dir, revision_folder)): continue for res_file in os.listdir(os.path.join(results_model_dir, revision_folder)): if (res_file.endswith(".json")) and not ( res_file.endswith(("overall_results.json", "model_meta.json")) ): results_model_file = os.path.join(results_model_dir, revision_folder, res_file) files[model_name].append(results_model_file) with open("paths.json", "w") as f: json.dump(files, f, indent=2) return files class MTEBResults(datasets.GeneratorBasedBuilder): """MTEBResults""" BUILDER_CONFIGS = [ datasets.BuilderConfig( name=model, description=f"{model} MTEB results", version=VERSION, ) for model in MODELS ] def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features( { "mteb_dataset_name": datasets.Value("string"), "eval_language": datasets.Value("string"), "metric": datasets.Value("string"), "score": datasets.Value("float"), "split": datasets.Value("string"), "hf_subset": datasets.Value("string"), } ), supervised_keys=None, citation=_CITATION, ) def _split_generators(self, dl_manager): path_file = dl_manager.download_and_extract(URL) # Local debugging help # with open("/path/to/local/paths.json") as f: with open(path_file) as f: files = json.load(f) downloaded_files = dl_manager.download_and_extract(files[self.config.name]) return [datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files})] def _generate_examples(self, filepath): """This function returns the examples in the raw (text) form.""" logger.info(f"Generating examples from {filepath}") out = [] for path in filepath: with open(path, encoding="utf-8") as f: res_dict = json.load(f) # Naming changed from mteb_dataset_name to task_name ds_name = res_dict.get("mteb_dataset_name", res_dict.get("task_name")) # New MTEB format uses scores res_dict = res_dict.get("scores", res_dict) split = "test" if (ds_name in TRAIN_SPLIT) and ("train" in res_dict): split = "train" elif (ds_name in VALIDATION_SPLIT) and ("validation" in res_dict): split = "validation" elif (ds_name in DEV_SPLIT) and ("dev" in res_dict): split = "dev" elif (ds_name in TESTFULL_SPLIT) and ("test.full" in res_dict): split = "test.full" elif ds_name in STANDARD_SPLIT: split = [] if "standard" in res_dict: split += ["standard"] if "long" in res_dict: split += ["long"] elif (ds_name in DEVTEST_SPLIT) and ("devtest" in res_dict): split = "devtest" elif ds_name in TEST_AVG_SPLIT: # Average splits res_dict = {} for split in TEST_AVG_SPLIT[ds_name]: # Old MTEB format if isinstance(res_dict.get(split), dict): for k, v in res_dict.get(split, {}).items(): if k in ["hf_subset", "languages"]: res_dict[k] = v v /= len(TEST_AVG_SPLIT[ds_name]) if k not in res_dict: res_dict[k] = v else: res_dict[k] += v # New MTEB format elif isinstance(res_dict.get(split), list): assert len(res_dict[split]) == 1, "Only single-lists supported for now" for k, v in res_dict[split][0].items(): if k in ["hf_subset", "languages"]: res_dict[k] = v if not isinstance(v, float): continue v /= len(TEST_AVG_SPLIT[ds_name]) if k not in res_dict: res_dict[k] = v else: res_dict[k] += v split = "test_avg" res_dict = {split: [res_dict]} elif "test" not in res_dict: print(f"Skipping {ds_name} as split {split} not present.") continue splits = [split] if not isinstance(split, list) else split full_res_dict = res_dict for split in splits: res_dict = full_res_dict.get(split) ### New MTEB format ### if isinstance(res_dict, list): for res in res_dict: lang = res.pop("languages", [""]) subset = res.pop("hf_subset", "") if len(lang) == 1: lang = lang[0].replace("eng-Latn", "") else: lang = "_".join(lang) if not lang: lang = subset for metric, score in res.items(): if metric in SKIP_KEYS: continue if isinstance(score, dict): # Legacy format with e.g. {cosine: {spearman: ...}} # Now it is {cosine_spearman: ...} for k, v in score.items(): if not isinstance(v, float): print(f"WARNING: Expected float, got {v} for {ds_name} {lang} {metric} {k}") continue if metric in SKIP_KEYS: continue out.append( { "mteb_dataset_name": ds_name, "eval_language": lang, "metric": metric + "_" + k, "score": v * 100, "hf_subset": subset, } ) else: if not isinstance(score, float): print(f"WARNING: Expected float, got {score} for {ds_name} {lang} {metric}") continue out.append( { "mteb_dataset_name": ds_name, "eval_language": lang, "metric": metric, "score": score * 100, "split": split, "hf_subset": subset, } ) ### Old MTEB format ### else: is_multilingual = any(x in res_dict for x in EVAL_LANGS) langs = res_dict.keys() if is_multilingual else ["en"] for lang in langs: if lang in SKIP_KEYS: continue test_result_lang = res_dict.get(lang) if is_multilingual else res_dict subset = test_result_lang.pop("hf_subset", "") if subset == "" and is_multilingual: subset = lang for metric, score in test_result_lang.items(): if not isinstance(score, dict): score = {metric: score} for sub_metric, sub_score in score.items(): if any(x in sub_metric for x in SKIP_KEYS): continue if isinstance(sub_score, dict): continue out.append( { "mteb_dataset_name": ds_name, "eval_language": lang if is_multilingual else "", "metric": f"{metric}_{sub_metric}" if metric != sub_metric else metric, "score": sub_score * 100, "split": split, "hf_subset": subset, } ) for idx, row in enumerate(sorted(out, key=lambda x: x["mteb_dataset_name"])): yield idx, row # NOTE: for generating the new paths if __name__ == "__main__": get_paths()