mteb 2.7.3__py3-none-any.whl → 2.7.5__py3-none-any.whl

mteb/models/model_implementations/bm25.py

@@ -13,7 +13,6 @@ if TYPE_CHECKING:
     from mteb.types import (
         CorpusDatasetType,
         EncodeKwargs,
-        InstructionDatasetType,
         QueryDatasetType,
         RetrievalOutputType,
         TopRankedDocumentsType,
@@ -55,6 +54,7 @@ def bm25_loader(model_name, **kwargs) -> SearchProtocol:
             hf_split: str,
             hf_subset: str,
             encode_kwargs: EncodeKwargs,
+            num_proc: int = 1,
         ) -> None:
             logger.info("Encoding Corpus...")
             corpus_texts = [
@@ -80,8 +80,8 @@ def bm25_loader(model_name, **kwargs) -> SearchProtocol:
             hf_subset: str,
             top_k: int,
             encode_kwargs: EncodeKwargs,
-            instructions: InstructionDatasetType | None = None,
             top_ranked: TopRankedDocumentsType | None = None,
+            num_proc: int = 1,
         ) -> RetrievalOutputType:
             logger.info("Encoding Queries...")
             query_ids = list(queries["id"])
@@ -103,13 +103,17 @@ def bm25_loader(model_name, **kwargs) -> SearchProtocol:
                 query_results = queries_results[qi]
                 scores = queries_scores[qi]
                 doc_id_to_score = {}
+                query_documents = (
+                    top_ranked[qid] if top_ranked and qid in top_ranked else None
+                )
 
                 # Iterate over results
-                for ri in range(len(query_results)):
-                    doc_idx = query_results[ri]
-                    score = scores[ri]
+                for doc_idx, score in zip(query_results, scores):
                     doc_id = self.corpus_idx_to_id[doc_idx]
 
+                    # handle reranking with a filtered set of documents
+                    if query_documents is not None and doc_id not in query_documents:
+                        continue
                     doc_id_to_score[doc_id] = float(score)
 
                 results[qid] = doc_id_to_score
@@ -132,6 +136,7 @@ bm25_s = ModelMeta(
     revision="0_1_10",
     release_date="2024-07-10",  # release of version 0.1.10
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     embed_dim=None,
     license=None,

mteb/models/model_implementations/bmretriever_models.py

@@ -103,6 +103,7 @@ BMRetriever_410M = ModelMeta(
     release_date="2024-04-29",
     embed_dim=1024,
     n_parameters=353_822_720,
+    n_embedding_parameters=51_511_296,
     memory_usage_mb=1349,
     max_tokens=2048,
     license="mit",
@@ -133,6 +134,7 @@ BMRetriever_1B = ModelMeta(
     release_date="2024-04-29",
     embed_dim=2048,
     n_parameters=908_759_040,
+    n_embedding_parameters=103_022_592,
     memory_usage_mb=3466,
     max_tokens=2048,
     license="mit",
@@ -163,6 +165,7 @@ BMRetriever_2B = ModelMeta(
     release_date="2024-04-29",
     embed_dim=2048,
     n_parameters=2_506_172_416,
+    n_embedding_parameters=524_288_000,
     memory_usage_mb=9560,
     max_tokens=8192,
     license="mit",
@@ -193,6 +196,7 @@ BMRetriever_7B = ModelMeta(
     release_date="2024-04-29",
     embed_dim=4096,
     n_parameters=7_110_660_096,
+    n_embedding_parameters=131_072_000,
     memory_usage_mb=27124,
     max_tokens=32768,
     license="mit",

mteb/models/model_implementations/cadet_models.py

@@ -41,6 +41,7 @@ cadet_embed = ModelMeta(
     open_weights=True,
     release_date="2025-05-11",
     n_parameters=109_000_000,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=418,
     embed_dim=768,
     license="apache-2.0",

mteb/models/model_implementations/cde_models.py

@@ -226,6 +226,7 @@ cde_small_v1 = ModelMeta(
     revision="e151df18af0d7f1d1c37b074fee58406ececf19f",
     release_date="2024-09-24",
     n_parameters=int(281 * 1e6),
+    n_embedding_parameters=None,
     memory_usage_mb=1072,  # Though the second-stage model is only 140M
     max_tokens=512,
     embed_dim=768,
@@ -255,6 +256,7 @@ cde_small_v2 = ModelMeta(
     revision="4e1d021a6c3fd7ce8aa0a7204057eee5ae61d390",
     release_date="2025-01-13",
     n_parameters=int(306 * 1e6),
+    n_embedding_parameters=None,
     memory_usage_mb=1166,  # Though the second-stage model is only 140M
     max_tokens=512,
     embed_dim=768,

mteb/models/model_implementations/clip_models.py

@@ -128,6 +128,7 @@ clip_vit_large_patch14 = ModelMeta(
     release_date="2021-02-26",
     modalities=["image", "text"],
     n_parameters=428_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1631,
     max_tokens=77,
     embed_dim=768,
@@ -152,6 +153,7 @@ clip_vit_base_patch32 = ModelMeta(
     release_date="2021-02-26",
     modalities=["image", "text"],
     n_parameters=151_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=576,
     max_tokens=77,
     embed_dim=512,
@@ -176,6 +178,7 @@ clip_vit_base_patch16 = ModelMeta(
     release_date="2021-02-26",
     modalities=["image", "text"],
     n_parameters=151_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=576,
     max_tokens=77,
     embed_dim=512,

mteb/models/model_implementations/clips_models.py

@@ -30,6 +30,7 @@ e5_nl_small = ModelMeta(
     revision="0243664a6c5e12eef854b091eb283e51833c3e9f",
     release_date="2025-09-23",
     n_parameters=40_800_000,
+    n_embedding_parameters=19_200_768,
     memory_usage_mb=78,
     embed_dim=384,
     license="mit",
@@ -57,6 +58,7 @@ e5_nl_base = ModelMeta(
     revision="6bd5722f236da48b4b8bcb28cc1fc478f7089956",
     release_date="2025-09-23",
     n_parameters=124_400_000,
+    n_embedding_parameters=38_401_536,
     memory_usage_mb=237,
     embed_dim=768,
     license="mit",
@@ -84,6 +86,7 @@ e5_nl_large = ModelMeta(
     revision="683333f86ed9eb3699b5567f0fdabeb958d412b0",
     release_date="2025-09-23",
     n_parameters=355_000_000,
+    n_embedding_parameters=51_202_048,
     memory_usage_mb=1355,
     embed_dim=1024,
     license="mit",

mteb/models/model_implementations/codefuse_models.py

@@ -236,6 +236,7 @@ F2LLM_0B6 = ModelMeta(
     revision="36416618b83d4bd84a8ca30c2ee01ed518f9f2e7",
     release_date="2025-09-18",
     n_parameters=595_776_512,
+    n_embedding_parameters=None,
     memory_usage_mb=1137,
     embed_dim=1024,
     license="apache-2.0",
@@ -266,6 +267,7 @@ F2LLM_1B7 = ModelMeta(
     revision="fdce0e09655f42cea26f7f66f5a70cd4507ea45c",
     release_date="2025-09-18",
     n_parameters=1_720_574_976,
+    n_embedding_parameters=None,
     memory_usage_mb=3282,
     embed_dim=2560,
     license="apache-2.0",
@@ -296,6 +298,7 @@ F2LLM_4B = ModelMeta(
     revision="9fe95901ed2b6b59dd7673d6e93c9d76766a1e25",
     release_date="2025-09-18",
     n_parameters=4_021_774_336,
+    n_embedding_parameters=None,
     memory_usage_mb=7672,
     embed_dim=2560,
     license="apache-2.0",
@@ -318,6 +321,7 @@ C2LLM_0B5 = ModelMeta(
     release_date="2025-12-22",
     languages=c2llm_languages,
     n_parameters=497252096,
+    n_embedding_parameters=None,
     memory_usage_mb=948.0,
     max_tokens=32768,
     embed_dim=896,
@@ -346,6 +350,7 @@ C2LLM_7B = ModelMeta(
     release_date="2025-12-22",
     languages=c2llm_languages,
     n_parameters=7667028992,
+    n_embedding_parameters=None,
     memory_usage_mb=14624.0,
     max_tokens=32768,
     embed_dim=3584,

mteb/models/model_implementations/codesage_models.py

@@ -28,6 +28,7 @@ codesage_large = ModelMeta(
     release_date="2024-02-03",
     modalities=["text"],
     n_parameters=1_300_000_000,
+    n_embedding_parameters=100_667_392,
     memory_usage_mb=4959,
     max_tokens=2048,
     embed_dim=2048,
@@ -55,6 +56,7 @@ codesage_base = ModelMeta(
     release_date="2024-02-03",
     modalities=["text"],
     n_parameters=356_000_000,
+    n_embedding_parameters=50_333_696,
     memory_usage_mb=1358,
     max_tokens=2048,
     embed_dim=1024,
@@ -82,6 +84,7 @@ codesage_small = ModelMeta(
     release_date="2024-02-03",
     modalities=["text"],
     n_parameters=130_000_000,
+    n_embedding_parameters=50_333_696,
     memory_usage_mb=496,
     max_tokens=2048,
     embed_dim=1024,

mteb/models/model_implementations/cohere_models.py

@@ -392,6 +392,7 @@ cohere_mult_3 = ModelMeta(
     revision="1",
     release_date="2023-11-02",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=None,
     embed_dim=512,
@@ -418,6 +419,7 @@ cohere_eng_3 = ModelMeta(
     revision="1",
     release_date="2023-11-02",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=512,
     embed_dim=1024,
@@ -443,6 +445,7 @@ cohere_mult_light_3 = ModelMeta(
     reference="https://cohere.com/blog/introducing-embed-v3",
     release_date="2023-11-02",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=512,
     embed_dim=384,
@@ -468,6 +471,7 @@ cohere_eng_light_3 = ModelMeta(
     revision="1",
     release_date="2023-11-02",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=512,
     embed_dim=384,

mteb/models/model_implementations/cohere_v.py

@@ -391,6 +391,7 @@ cohere_mult_3 = ModelMeta(
     revision="1",
     release_date="2024-10-24",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=None,
     embed_dim=1024,
@@ -415,6 +416,7 @@ cohere_eng_3 = ModelMeta(
     revision="1",
     release_date="2024-10-24",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=None,
     embed_dim=1024,
@@ -439,6 +441,7 @@ cohere_embed_v4_multimodal = ModelMeta(
     revision="1",
     release_date="2024-12-01",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=128000,
     embed_dim=1536,
@@ -463,6 +466,7 @@ cohere_embed_v4_multimodal_binary = ModelMeta(
     revision="1",
     release_date="2024-12-01",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=128000,
     embed_dim=1536,
@@ -488,6 +492,7 @@ cohere_embed_v4_multimodal_int8 = ModelMeta(
     revision="1",
     release_date="2024-12-01",
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=128000,
     embed_dim=1536,

mteb/models/model_implementations/colpali_models.py

@@ -220,6 +220,7 @@ colpali_v1_1 = ModelMeta(
     release_date="2024-08-21",
     modalities=["image", "text"],
     n_parameters=2_920_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4700,
     max_tokens=16384,
     embed_dim=128,
@@ -247,6 +248,7 @@ colpali_v1_2 = ModelMeta(
     release_date="2024-08-26",
     modalities=["image", "text"],
     n_parameters=2_920_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4700,
     max_tokens=16384,
     embed_dim=128,
@@ -274,6 +276,7 @@ colpali_v1_3 = ModelMeta(
     release_date="2024-11-01",
     modalities=["image", "text"],
     n_parameters=2_920_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4700,
     max_tokens=16384,
     embed_dim=128,

mteb/models/model_implementations/colqwen_models.py

@@ -224,6 +224,7 @@ colqwen2 = ModelMeta(
     release_date="2025-11-03",
     modalities=["image", "text"],
     n_parameters=2_210_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7200,
     max_tokens=32768,
     embed_dim=128,
@@ -251,6 +252,7 @@ colqwen2_5 = ModelMeta(
     release_date="2025-01-31",
     modalities=["image", "text"],
     n_parameters=3_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7200,
     max_tokens=128000,
     embed_dim=128,
@@ -295,6 +297,7 @@ colqwen3_8b = ModelMeta(
     release_date="2025-11-26",
     modalities=["image", "text"],
     n_parameters=8_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=16724,
     max_tokens=262144,
     embed_dim=320,
@@ -319,6 +322,7 @@ colqwen3_4b = ModelMeta(
     release_date="2025-11-26",
     modalities=["image", "text"],
     n_parameters=4_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=8466,
     max_tokens=262144,
     embed_dim=320,
@@ -365,6 +369,7 @@ colnomic_3b = ModelMeta(
     release_date="2025-03-31",
     modalities=["image", "text"],
     n_parameters=3_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7200,
     max_tokens=128000,
     embed_dim=128,
@@ -430,6 +435,7 @@ evoqwen25_vl_retriever_3b_v1 = ModelMeta(
     release_date="2025-11-04",
     modalities=["image", "text"],
     n_parameters=3_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7200,
     max_tokens=128000,
     embed_dim=128,
@@ -456,6 +462,7 @@ evoqwen25_vl_retriever_7b_v1 = ModelMeta(
     release_date="2025-11-04",
     modalities=["image", "text"],
     n_parameters=7_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=14400,
     max_tokens=128000,
     embed_dim=128,

mteb/models/model_implementations/colsmol_models.py

@@ -60,6 +60,7 @@ colsmol_256m = ModelMeta(
     release_date="2025-01-22",
     modalities=["image", "text"],
     n_parameters=256_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=800,
     max_tokens=8192,
     embed_dim=128,
@@ -87,6 +88,7 @@ colsmol_500m = ModelMeta(
     release_date="2025-01-22",
     modalities=["image", "text"],
     n_parameters=500_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1200,
     max_tokens=8192,
     embed_dim=128,

mteb/models/model_implementations/conan_models.py

@@ -211,6 +211,7 @@ Conan_embedding_v2 = ModelMeta(
     embed_dim=3584,
     open_weights=False,
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     license="apache-2.0",
     reference="https://huggingface.co/TencentBAC/Conan-embedding-v2",

mteb/models/model_implementations/dino_models.py

@@ -117,6 +117,7 @@ dinov2_small = ModelMeta(
     release_date="2023-07-18",
     modalities=["image"],
     n_parameters=22_100_000,
+    n_embedding_parameters=None,
     memory_usage_mb=84,
     max_tokens=None,
     embed_dim=384,
@@ -148,6 +149,7 @@ dinov2_base = ModelMeta(
     release_date="2023-07-18",
     modalities=["image"],
     n_parameters=86_600_000,
+    n_embedding_parameters=None,
     memory_usage_mb=330,
     max_tokens=None,
     embed_dim=768,
@@ -179,6 +181,7 @@ dinov2_large = ModelMeta(
     release_date="2023-07-18",
     modalities=["image"],
     n_parameters=304_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1161,
     max_tokens=None,
     embed_dim=1024,
@@ -210,6 +213,7 @@ dinov2_giant = ModelMeta(
     release_date="2023-07-18",
     modalities=["image"],
     n_parameters=1_140_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4335,
     max_tokens=None,
     embed_dim=1536,
@@ -245,6 +249,7 @@ webssl_dino300m_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=304_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1158,
     max_tokens=None,
     embed_dim=1024,
@@ -276,6 +281,7 @@ webssl_dino1b_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=1_130_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4329,
     max_tokens=None,
     embed_dim=1536,
@@ -307,6 +313,7 @@ webssl_dino2b_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=2_080_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7951,
     max_tokens=None,
     embed_dim=2688,
@@ -338,6 +345,7 @@ webssl_dino3b_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=3_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=11247,
     max_tokens=None,
     embed_dim=3072,
@@ -369,6 +377,7 @@ webssl_dino5b_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=5_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=18838,
     max_tokens=None,
     embed_dim=3584,
@@ -400,6 +409,7 @@ webssl_dino7b_full8b_224 = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=7_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=24605,
     max_tokens=None,
     embed_dim=4096,
@@ -431,6 +441,7 @@ webssl_dino7b_full8b_378 = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=7_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=24613,
     max_tokens=None,
     embed_dim=4096,
@@ -462,6 +473,7 @@ webssl_dino7b_full8b_518 = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=7_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=24623,
     max_tokens=None,
     embed_dim=4096,
@@ -494,6 +506,7 @@ webssl_dino2b_light2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=2_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7951,
     max_tokens=None,
     embed_dim=2688,
@@ -525,6 +538,7 @@ webssl_dino2b_heavy2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=2_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=7951,
     max_tokens=None,
     embed_dim=2688,
@@ -556,6 +570,7 @@ webssl_dino3b_light2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=3_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=11247,
     max_tokens=None,
     embed_dim=3072,
@@ -587,6 +602,7 @@ webssl_dino3b_heavy2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=3_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=11247,
     max_tokens=None,
     embed_dim=3072,
@@ -618,6 +634,7 @@ webssl_mae300m_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=304_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1161,
     max_tokens=None,
     embed_dim=1024,
@@ -649,6 +666,7 @@ webssl_mae700m_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=700_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=2412,
     max_tokens=None,
     embed_dim=1280,
@@ -680,6 +698,7 @@ webssl_mae1b_full2b = ModelMeta(
     release_date="2025-04-24",
     modalities=["image"],
     n_parameters=1_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4337,
     max_tokens=None,
     embed_dim=1536,

mteb/models/model_implementations/e5_instruct.py

@@ -57,6 +57,7 @@ e5_instruct = ModelMeta(
     use_instructions=True,
     reference="https://huggingface.co/intfloat/multilingual-e5-large-instruct",
     n_parameters=560_000_000,
+    n_embedding_parameters=256_002_048,
     memory_usage_mb=1068,
     embed_dim=1024,
     license="mit",
@@ -102,6 +103,7 @@ e5_mistral = ModelMeta(
     use_instructions=True,
     reference="https://huggingface.co/intfloat/e5-mistral-7b-instruct",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=131_072_000,
     memory_usage_mb=13563,
     embed_dim=4096,
     license="mit",
@@ -145,6 +147,7 @@ zeta_alpha_ai__zeta_alpha_e5_mistral = ModelMeta(
     release_date="2024-08-30",
     languages=["eng-Latn"],
     n_parameters=7110660096,
+    n_embedding_parameters=None,
     memory_usage_mb=13563,
     max_tokens=32768.0,
     embed_dim=4096,
@@ -228,6 +231,7 @@ BeastyZ__e5_R_mistral_7b = ModelMeta(
     release_date="2024-06-28",
     languages=["eng-Latn"],
     n_parameters=7241732096,
+    n_embedding_parameters=131_072_000,
     memory_usage_mb=27625,
     max_tokens=32768.0,
     embed_dim=4096,

mteb/models/model_implementations/e5_models.py

@@ -76,6 +76,7 @@ e5_mult_small = ModelMeta(
     revision="fd1525a9fd15316a2d503bf26ab031a61d056e98",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=118_000_000,
+    n_embedding_parameters=96_014_208,
     memory_usage_mb=449,
     embed_dim=384,
     license="mit",
@@ -103,6 +104,7 @@ e5_mult_base = ModelMeta(
     revision="d13f1b27baf31030b7fd040960d60d909913633f",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=278_000_000,
+    n_embedding_parameters=192_001_536,
     memory_usage_mb=1061,
     embed_dim=768,
     license="mit",
@@ -130,6 +132,7 @@ e5_mult_large = ModelMeta(
     revision="ab10c1a7f42e74530fe7ae5be82e6d4f11a719eb",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=560_000_000,
+    n_embedding_parameters=256_002_048,
     memory_usage_mb=2136,
     embed_dim=1024,
     license="mit",
@@ -157,6 +160,7 @@ e5_eng_small_v2 = ModelMeta(
     revision="dca8b1a9dae0d4575df2bf423a5edb485a431236",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=33_000_000,
+    n_embedding_parameters=11_720_448,
     memory_usage_mb=127,
     embed_dim=384,
     license="mit",
@@ -184,6 +188,7 @@ e5_eng_small = ModelMeta(
     revision="e272f3049e853b47cb5ca3952268c6662abda68f",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=33_000_000,
+    n_embedding_parameters=11_720_448,
     memory_usage_mb=127,
     embed_dim=384,
     license="mit",
@@ -211,6 +216,7 @@ e5_eng_base_v2 = ModelMeta(
     revision="1c644c92ad3ba1efdad3f1451a637716616a20e8",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=109_000_000,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=418,
     embed_dim=768,
     license="mit",
@@ -239,6 +245,7 @@ e5_eng_large_v2 = ModelMeta(
     revision="b322e09026e4ea05f42beadf4d661fb4e101d311",
     release_date=E5_PAPER_RELEASE_DATE,
     n_parameters=335_000_000,
+    n_embedding_parameters=31_254_528,
     memory_usage_mb=1278,
     embed_dim=1024,
     license="mit",
@@ -267,6 +274,7 @@ e5_large = ModelMeta(
     revision="4dc6d853a804b9c8886ede6dda8a073b7dc08a81",
     release_date="2022-12-26",
     n_parameters=335_000_000,
+    n_embedding_parameters=31_254_528,
     memory_usage_mb=1278,
     embed_dim=1024,
     license="apache-2.0",
@@ -295,6 +303,7 @@ e5_base = ModelMeta(
     revision="b533fe4636f4a2507c08ddab40644d20b0006d6a",
     release_date="2022-12-26",
     n_parameters=109_000_000,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=418,
     embed_dim=768,
     license="apache-2.0",

mteb/models/model_implementations/e5_v.py

@@ -171,6 +171,7 @@ e5_v = ModelMeta(
     release_date="2024-07-17",
     modalities=["image", "text"],
     n_parameters=8_360_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=15936,
     max_tokens=8192,
     embed_dim=4096,

mteb/models/model_implementations/eagerworks_models.py

@@ -153,6 +153,7 @@ Eager_Embed_V1 = ModelMeta(
     release_date="2025-11-20",
     modalities=["image", "text"],
     n_parameters=4_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=16929,
     max_tokens=262144,
     embed_dim=2560,

mteb/models/model_implementations/emillykkejensen_models.py

@@ -10,6 +10,7 @@ embedding_gemma_300m_scandi = ModelMeta(
     revision="9f3307b9f601db564a9190cb475324d128dcfe86",
     release_date="2025-10-17",
     n_parameters=307_581_696,
+    n_embedding_parameters=None,
     embed_dim=768,
     max_tokens=2048,
     license="apache-2.0",
@@ -43,6 +44,7 @@ qwen_scandi = ModelMeta(
     revision="cf1e7ba36ebd3d605549d8f02930a18e17b54513",
     release_date="2025-10-17",
     n_parameters=595776512,
+    n_embedding_parameters=None,
     memory_usage_mb=2272,
     embed_dim=1024,
     max_tokens=32768,
@@ -67,6 +69,7 @@ mmbert_scandi = ModelMeta(
     revision="82d74c7a5d8e1ddf31b132865df2d16b2b0294ee",
     release_date="2025-10-17",
     n_parameters=306939648,
+    n_embedding_parameters=None,
     memory_usage_mb=1171,
     embed_dim=768,
     max_tokens=8192,