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

mteb/models/model_implementations/model2vec_models.py

@@ -173,6 +173,7 @@ m2v_base_glove_subword = ModelMeta(
     revision="5f4f5ca159b7321a8b39739bba0794fa0debddf4",
     release_date="2024-09-21",
     n_parameters=int(103 * 1e6),
+    n_embedding_parameters=int(103 * 1e6),
     memory_usage_mb=391,
     max_tokens=np.inf,  # Theoretically infinite
     embed_dim=256,
@@ -199,6 +200,7 @@ m2v_base_glove = ModelMeta(
     revision="38ebd7f10f71e67fa8db898290f92b82e9cfff2b",
     release_date="2024-09-21",
     n_parameters=int(102 * 1e6),
+    n_embedding_parameters=int(102 * 1e6),
     memory_usage_mb=391,
     max_tokens=np.inf,
     embed_dim=256,
@@ -224,6 +226,7 @@ m2v_base_output = ModelMeta(
     revision="02460ae401a22b09d2c6652e23371398329551e2",
     release_date="2024-09-21",
     n_parameters=int(7.56 * 1e6),
+    n_embedding_parameters=int(7.56 * 1e6),
     memory_usage_mb=29,
     max_tokens=np.inf,
     embed_dim=256,
@@ -249,6 +252,7 @@ m2v_multilingual_output = ModelMeta(
     revision="2cf4ec4e1f51aeca6c55cf9b93097d00711a6305",
     release_date="2024-09-21",
     n_parameters=int(128 * 1e6),
+    n_embedding_parameters=int(128 * 1e6),
     memory_usage_mb=489,
     max_tokens=np.inf,
     embed_dim=256,
@@ -274,6 +278,7 @@ potion_base_2m = ModelMeta(
     revision="86db093558fbced2072b929eb1690bce5272bd4b",
     release_date="2024-10-29",
     n_parameters=int(2 * 1e6),
+    n_embedding_parameters=int(2 * 1e6),
     memory_usage_mb=7,
     max_tokens=np.inf,
     embed_dim=64,
@@ -299,6 +304,7 @@ potion_base_4m = ModelMeta(
     revision="81b1802ada41afcd0987a37dc15e569c9fa76f04",
     release_date="2024-10-29",
     n_parameters=int(3.78 * 1e6),
+    n_embedding_parameters=int(3.78 * 1e6),
     memory_usage_mb=14,
     max_tokens=np.inf,
     embed_dim=128,
@@ -324,6 +330,7 @@ potion_base_8m = ModelMeta(
     revision="dcbec7aa2d52fc76754ac6291803feedd8c619ce",
     release_date="2024-10-29",
     n_parameters=int(7.56 * 1e6),
+    n_embedding_parameters=int(7.56 * 1e6),
     memory_usage_mb=29,
     max_tokens=np.inf,
     embed_dim=256,
@@ -349,6 +356,7 @@ potion_multilingual_128m = ModelMeta(
     revision="38ebd7f10f71e67fa8db898290f92b82e9cfff2a",
     release_date="2025-05-23",
     n_parameters=128 * 1e6,
+    n_embedding_parameters=128 * 1e6,
     memory_usage_mb=489,
     max_tokens=np.inf,
     embed_dim=256,
@@ -374,6 +382,7 @@ pubmed_bert_100k = ModelMeta(
     revision="bac5e3b12fb8c650e92a19c41b436732c4f16e9e",
     release_date="2025-01-03",
     n_parameters=1 * 1e5,
+    n_embedding_parameters=1 * 1e5,
     memory_usage_mb=0,
     max_tokens=np.inf,
     embed_dim=64,
@@ -398,6 +407,7 @@ pubmed_bert_500k = ModelMeta(
     revision="34ba71e35c393fdad7ed695113f653feb407b16b",
     release_date="2025-01-03",
     n_parameters=5 * 1e5,
+    n_embedding_parameters=5 * 1e5,
     memory_usage_mb=2,
     max_tokens=np.inf,
     embed_dim=64,
@@ -422,6 +432,7 @@ pubmed_bert_1m = ModelMeta(
     revision="2b7fed222594708da6d88bcda92ae9b434b7ddd1",
     release_date="2025-01-03",
     n_parameters=1 * 1e6,
+    n_embedding_parameters=1 * 1e6,
     memory_usage_mb=2,
     max_tokens=np.inf,
     embed_dim=64,
@@ -446,6 +457,7 @@ pubmed_bert_2m = ModelMeta(
     revision="1d7bbe04d6713e425161146bfdc71473cbed498a",
     release_date="2025-01-03",
     n_parameters=1.95 * 1e6,
+    n_embedding_parameters=1.95 * 1e6,
     memory_usage_mb=7,
     max_tokens=np.inf,
     embed_dim=64,
@@ -470,6 +482,7 @@ pubmed_bert_8m = ModelMeta(
     revision="387d350015e963744f4fafe56a574b7cd48646c9",
     release_date="2025-01-03",
     n_parameters=7.81 * 1e6,
+    n_embedding_parameters=7.81 * 1e6,
     memory_usage_mb=30,
     max_tokens=np.inf,
     embed_dim=256,

mteb/models/model_implementations/moka_models.py

@@ -97,6 +97,7 @@ m3e_base = ModelMeta(
     revision="764b537a0e50e5c7d64db883f2d2e051cbe3c64c",
     release_date="2023-06-06",  # first commit
     n_parameters=int(102 * 1e6),
+    n_embedding_parameters=16_226_304,
     memory_usage_mb=390,
     embed_dim=768,
     # They don't give a specific license but commercial use is not allowed
@@ -123,6 +124,7 @@ m3e_small = ModelMeta(
     revision="44c696631b2a8c200220aaaad5f987f096e986df",
     release_date="2023-06-02",  # first commit
     n_parameters=None,
+    n_embedding_parameters=10_817_536,
     memory_usage_mb=None,  # Can't be seen on HF page
     embed_dim=512,
     # They don't give a specific license but commercial use is not allowed
@@ -149,6 +151,7 @@ m3e_large = ModelMeta(
     revision="12900375086c37ba5d83d1e417b21dc7d1d1f388",
     release_date="2023-06-21",  # first commit
     n_parameters=None,
+    n_embedding_parameters=21_635_072,
     memory_usage_mb=None,  # Can't be seen on HF page
     embed_dim=768,
     # They don't give a specific license but commercial use is not allowed

mteb/models/model_implementations/nbailab.py

@@ -12,6 +12,7 @@ nb_sbert = ModelMeta(
     revision="b95656350a076aeafd2d23763660f80655408cc6",
     release_date="2022-11-23",
     n_parameters=1_780_000_000,
+    n_embedding_parameters=91_812_096,
     memory_usage_mb=678,
     embed_dim=4096,
     license="apache-2.0",
@@ -34,6 +35,7 @@ nb_bert_large = ModelMeta(
     revision="f9d0fc184adab4dc354d85e1854b7634540d7550",
     release_date="2021-04-29",
     n_parameters=355087360,
+    n_embedding_parameters=51_200_000,
     memory_usage_mb=1359,
     embed_dim=1024,
     license="cc-by-4.0",
@@ -56,6 +58,7 @@ nb_bert_base = ModelMeta(
     revision="9417c3f62a3adc99f17ff92bff446f35d011f994",
     release_date="2021-01-13",
     n_parameters=177853440,
+    n_embedding_parameters=91_812_096,
     memory_usage_mb=681,
     embed_dim=768,
     license="cc-by-4.0",

mteb/models/model_implementations/no_instruct_sentence_models.py

@@ -110,6 +110,7 @@ no_instruct_small_v0 = ModelMeta(
     revision="b38747000553d8268915c95a55fc87e707c9aadd",
     release_date="2024-05-01",  # first commit
     n_parameters=33_400_000,
+    n_embedding_parameters=11_720_448,
     memory_usage_mb=127,
     max_tokens=512,
     embed_dim=384,

mteb/models/model_implementations/nomic_models.py

@@ -215,6 +215,7 @@ nomic_embed_v1_5 = ModelMeta(
     release_date="2024-02-10",  # first commit
     citation=NOMIC_CITATION,
     n_parameters=137_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=522,
     max_tokens=8192,
     embed_dim=768,
@@ -249,6 +250,7 @@ nomic_embed_v1 = ModelMeta(
     revision="0759316f275aa0cb93a5b830973843ca66babcf5",
     release_date="2024-01-31",  # first commit
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=522,
     max_tokens=8192,
     embed_dim=768,
@@ -284,6 +286,7 @@ nomic_embed_v1_ablated = ModelMeta(
     revision="7d948905c5d5d3874fa55a925d68e49dbf411e5f",
     release_date="2024-01-15",  # first commit
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=8192,
     embed_dim=768,
@@ -312,6 +315,7 @@ nomic_embed_v1_unsupervised = ModelMeta(
     revision="b53d557b15ae63852847c222d336c1609eced93c",
     release_date="2024-01-15",  # first commit
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=8192,
     embed_dim=768,
@@ -340,6 +344,7 @@ nomic_modern_bert_embed = ModelMeta(
     revision="5960f1566fb7cb1adf1eb6e816639cf4646d9b12",
     release_date="2024-12-29",
     n_parameters=149_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=568,
     max_tokens=8192,
     embed_dim=768,
@@ -479,6 +484,7 @@ nomic_embed_text_v2_moe = ModelMeta(
     revision="1066b6599d099fbb93dfcb64f9c37a7c9e503e85",
     release_date="2025-02-07",
     n_parameters=475292928,
+    n_embedding_parameters=None,
     memory_usage_mb=1813,
     max_tokens=512,
     embed_dim=768,

mteb/models/model_implementations/nomic_models_vision.py

@@ -175,6 +175,7 @@ nomic_embed_vision_v1_5 = ModelMeta(
     release_date="2024-06-08",
     modalities=["image", "text"],
     n_parameters=92_900_000,
+    n_embedding_parameters=None,
     memory_usage_mb=355,
     max_tokens=2048,
     embed_dim=768,

mteb/models/model_implementations/nvidia_llama_nemoretriever_colemb.py

@@ -162,6 +162,7 @@ llama_nemoretriever_colembed_1b_v1 = ModelMeta(
     release_date="2025-06-27",
     modalities=["image", "text"],
     n_parameters=2_418_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=4610,
     max_tokens=8192,
     embed_dim=2048,
@@ -189,6 +190,7 @@ llama_nemoretriever_colembed_3b_v1 = ModelMeta(
     release_date="2025-06-27",
     modalities=["image", "text"],
     n_parameters=4_407_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=8403,
     max_tokens=8192,
     embed_dim=3072,

mteb/models/model_implementations/nvidia_models.py

@@ -204,6 +204,7 @@ NV_embed_v2 = ModelMeta(
     revision="7604d305b621f14095a1aa23d351674c2859553a",
     release_date="2024-09-09",  # initial commit of hf model.
     n_parameters=7_850_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=14975,
     embed_dim=4096,
     license="cc-by-nc-4.0",
@@ -235,6 +236,7 @@ NV_embed_v1 = ModelMeta(
     revision="570834afd5fef5bf3a3c2311a2b6e0a66f6f4f2c",
     release_date="2024-09-13",  # initial commit of hf model.
     n_parameters=7_850_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=14975,
     embed_dim=4096,
     license="cc-by-nc-4.0",
@@ -624,6 +626,7 @@ llama_embed_nemotron_8b = ModelMeta(
     revision="84a375593d27d3528beb4e104822515659e093b4",
     release_date="2025-10-23",
     n_parameters=7_504_924_672,
+    n_embedding_parameters=None,
     memory_usage_mb=28629,
     embed_dim=4096,
     license="https://huggingface.co/nvidia/llama-embed-nemotron-8b/blob/main/LICENSE",

mteb/models/model_implementations/octen_models.py

@@ -208,6 +208,7 @@ Octen_Embedding_4B = ModelMeta(
     revision="6e188e3b072c3e3678b235ad84e6e97bcbb71e8f",
     release_date="2025-12-30",
     n_parameters=4021774336,
+    n_embedding_parameters=None,
     memory_usage_mb=7671,
     embed_dim=2560,
     max_tokens=32768,
@@ -238,6 +239,7 @@ Octen_Embedding_8B = ModelMeta(
     revision="f7db178d5a82fb841f606a6a67c423cead2fdbba",
     release_date="2025-12-23",
     n_parameters=7567295488,
+    n_embedding_parameters=None,
     memory_usage_mb=14433,
     embed_dim=4096,
     max_tokens=32768,

mteb/models/model_implementations/openai_models.py

@@ -185,6 +185,7 @@ text_embedding_3_small = ModelMeta(
     embed_dim=1536,
     open_weights=False,
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     license=None,
     reference="https://openai.com/index/new-embedding-models-and-api-updates/",
@@ -213,6 +214,7 @@ text_embedding_3_large = ModelMeta(
     framework=["API"],
     use_instructions=False,
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     public_training_code=None,
     public_training_data=None,  # assumed
@@ -238,6 +240,7 @@ text_embedding_ada_002 = ModelMeta(
     framework=["API"],
     use_instructions=False,
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     public_training_code=None,
     public_training_data=None,  # assumed
@@ -262,6 +265,7 @@ text_embedding_3_small_512 = ModelMeta(
     embed_dim=512,
     open_weights=False,
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     license=None,
     reference="https://openai.com/index/new-embedding-models-and-api-updates/",
@@ -292,6 +296,7 @@ text_embedding_3_large_512 = ModelMeta(
     framework=["API"],
     use_instructions=False,
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     public_training_code=None,
     public_training_data=None,  # assumed

mteb/models/model_implementations/openclip_models.py

@@ -133,6 +133,7 @@ CLIP_ViT_L_14_DataComp_XL_s13B_b90K = ModelMeta(
     release_date="2023-04-26",
     modalities=["image", "text"],
     n_parameters=428_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1633,
     max_tokens=77,
     embed_dim=768,
@@ -159,6 +160,7 @@ CLIP_ViT_B_32_DataComp_XL_s13B_b90K = ModelMeta(
     release_date="2023-04-26",
     modalities=["image", "text"],
     n_parameters=151_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=576,
     max_tokens=77,
     embed_dim=512,
@@ -185,6 +187,7 @@ CLIP_ViT_B_16_DataComp_XL_s13B_b90K = ModelMeta(
     release_date="2023-04-26",
     modalities=["image", "text"],
     n_parameters=150_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=572,
     max_tokens=77,
     embed_dim=512,
@@ -211,6 +214,7 @@ CLIP_ViT_bigG_14_laion2B_39B_b160k = ModelMeta(
     release_date="2023-01-23",
     modalities=["image", "text"],
     n_parameters=2_540_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=9689,
     max_tokens=77,
     embed_dim=1280,
@@ -237,6 +241,7 @@ CLIP_ViT_g_14_laion2B_s34B_b88K = ModelMeta(
     release_date="2023-03-06",
     modalities=["image", "text"],
     n_parameters=1_367_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=5215,
     max_tokens=77,
     embed_dim=1024,
@@ -263,6 +268,7 @@ CLIP_ViT_H_14_laion2B_s32B_b79K = ModelMeta(
     release_date="2022-09-15",
     modalities=["image", "text"],
     n_parameters=986_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=3762,
     max_tokens=77,
     embed_dim=1024,
@@ -289,6 +295,7 @@ CLIP_ViT_L_14_laion2B_s32B_b82K = ModelMeta(
     release_date="2022-09-15",
     modalities=["image", "text"],
     n_parameters=428_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1631,
     max_tokens=77,
     embed_dim=768,
@@ -315,6 +322,7 @@ CLIP_ViT_B_32_laion2B_s34B_b79K = ModelMeta(
     release_date="2022-09-15",
     modalities=["image", "text"],
     n_parameters=151_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=577,
     max_tokens=77,
     embed_dim=512,

mteb/models/model_implementations/opensearch_neural_sparse_models.py

@@ -140,6 +140,7 @@ opensearch_neural_sparse_encoding_doc_v3_gte = ModelMeta(
     revision="a8abaa916125ee512a7a8f4d706d07eb0128a8e6",
     release_date="2025-06-18",
     n_parameters=137_394_234,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=549,
     embed_dim=30522,
     license="apache-2.0",
@@ -166,6 +167,7 @@ opensearch_neural_sparse_encoding_doc_v3_distill = ModelMeta(
     revision="babf71f3c48695e2e53a978208e8aba48335e3c0",
     release_date="2025-03-28",
     n_parameters=66_985_530,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=267,
     embed_dim=30522,
     license="apache-2.0",
@@ -188,6 +190,7 @@ opensearch_neural_sparse_encoding_doc_v2_distill = ModelMeta(
     revision="8921a26c78b8559d6604eb1f5c0b74c079bee38f",
     release_date="2024-07-17",
     n_parameters=66_985_530,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=267,
     embed_dim=30522,
     license="apache-2.0",
@@ -211,6 +214,7 @@ opensearch_neural_sparse_encoding_doc_v2_mini = ModelMeta(
     revision="4af867a426867dfdd744097531046f4289a32fdd",
     release_date="2024-07-18",
     n_parameters=22_744_506,
+    n_embedding_parameters=11_720_448,
     memory_usage_mb=86,
     embed_dim=30522,
     license="apache-2.0",
@@ -233,6 +237,7 @@ opensearch_neural_sparse_encoding_doc_v1 = ModelMeta(
     revision="98cdcbd72867c547f72f2b7b7bed9cdf9f09922d",
     release_date="2024-03-07",
     n_parameters=132_955_194,
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=507,
     embed_dim=30522,
     license="apache-2.0",

mteb/models/model_implementations/ops_moa_models.py

@@ -33,6 +33,7 @@ ops_moa_conan_embedding = ModelMeta(
     languages=["zho-Hans"],
     loader=OPSWrapper,
     n_parameters=int(343 * 1e6),
+    n_embedding_parameters=21_635_072,
     memory_usage_mb=1308,
     max_tokens=512,
     embed_dim=1536,
@@ -65,6 +66,7 @@ ops_moa_yuan_embedding = ModelMeta(
     languages=["zho-Hans"],
     loader=OPSWrapper,
     n_parameters=int(343 * 1e6),
+    n_embedding_parameters=21_635_072,
     memory_usage_mb=1242,
     max_tokens=512,
     embed_dim=1536,

mteb/models/model_implementations/ordalietech_solon_embeddings_mini_beta_1_1.py

@@ -4,6 +4,7 @@ solon_embeddings_1_1 = ModelMeta(
     name="OrdalieTech/Solon-embeddings-mini-beta-1.1",
     languages=["fra-Latn"],
     n_parameters=210_000_000,
+    n_embedding_parameters=None,
     public_training_code=None,
     memory_usage_mb=808.0,
     open_weights=True,

mteb/models/model_implementations/pawan_models.py

@@ -20,6 +20,7 @@ pawan_embd_68m = ModelMeta(
     revision="32f295145802bdbd65699ad65fd27d2a5b69a909",
     release_date="2025-12-08",
     n_parameters=68_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=260,
     embed_dim=768,
     license="apache-2.0",

mteb/models/model_implementations/piccolo_models.py

@@ -12,6 +12,7 @@ piccolo_base_zh = ModelMeta(
     revision="47c0a63b8f667c3482e05b2fd45577bb19252196",
     release_date="2023-09-04",  # first commit
     n_parameters=None,
+    n_embedding_parameters=16_226_304,
     memory_usage_mb=None,  # can't see on model card
     embed_dim=768,
     license="mit",
@@ -37,6 +38,7 @@ piccolo_large_zh_v2 = ModelMeta(
     revision="05948c1d889355936bdf9db7d30df57dd78d25a3",
     release_date="2024-04-22",  # first commit
     n_parameters=None,
+    n_embedding_parameters=None,
     memory_usage_mb=None,  # we don't know because they removed the model
     embed_dim=1024,
     license="not specified",

mteb/models/model_implementations/promptriever_models.py

@@ -87,6 +87,7 @@ promptriever_llama2 = ModelMeta(
     revision="01c7f73d771dfac7d292323805ebc428287df4f9-30b14e3813c0fa45facfd01a594580c3fe5ecf23",  # base-peft revision
     release_date="2024-09-15",
     n_parameters=7_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=26703,
     max_tokens=4096,
     embed_dim=4096,
@@ -123,6 +124,7 @@ promptriever_llama3 = ModelMeta(
     },
     release_date="2024-09-15",
     n_parameters=8_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=30518,
     max_tokens=8192,
     embed_dim=4096,
@@ -152,6 +154,7 @@ promptriever_llama3_instruct = ModelMeta(
     revision="5206a32e0bd3067aef1ce90f5528ade7d866253f-8b677258615625122c2eb7329292b8c402612c21",  # base-peft revision
     release_date="2024-09-15",
     n_parameters=8_000_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=30518,
     max_tokens=8192,
     embed_dim=4096,
@@ -185,6 +188,7 @@ promptriever_mistral_v1 = ModelMeta(
     revision="7231864981174d9bee8c7687c24c8344414eae6b-876d63e49b6115ecb6839893a56298fadee7e8f5",  # base-peft revision
     release_date="2024-09-15",
     n_parameters=7_000_000_000,
+    n_embedding_parameters=131_072_000,
     memory_usage_mb=26703,
     training_datasets={
         # "samaya-ai/msmarco-w-instructions",

mteb/models/model_implementations/pylate_models.py

@@ -53,6 +53,7 @@ class PylateSearchEncoder:
         hf_split: str,
         hf_subset: str,
         encode_kwargs: EncodeKwargs,
+        num_proc: int,
     ) -> None:
         """Index the corpus for retrieval.
 
@@ -62,6 +63,7 @@ class PylateSearchEncoder:
             hf_split: Split of current task, allows to know some additional information about current split.
             hf_subset: Subset of current task. Similar to `hf_split` to get more information
             encode_kwargs: Additional arguments to pass to the encoder during indexing.
+            num_proc: Number of processes to use for indexing.
         """
         self.task_corpus = corpus
 
@@ -87,12 +89,14 @@ class PylateSearchEncoder:
         top_k: int,
         encode_kwargs: EncodeKwargs,
         top_ranked: TopRankedDocumentsType | None = None,
+        num_proc: int,
     ) -> RetrievalOutputType:
         queries_dataloader = create_dataloader(
             queries,
             task_metadata,
             prompt_type=PromptType.query,
             batch_size=encode_kwargs.get("batch_size", 32),
+            num_proc=num_proc,
         )
 
         query_embeddings = self.encode(
@@ -116,6 +120,7 @@ class PylateSearchEncoder:
                 hf_subset=hf_subset,
                 hf_split=hf_split,
                 encode_kwargs=encode_kwargs,
+                num_proc=num_proc,
             )
         else:
             result_heaps = self._pylate_full_corpus_search(
@@ -126,6 +131,7 @@ class PylateSearchEncoder:
                 hf_subset=hf_subset,
                 hf_split=hf_split,
                 encode_kwargs=encode_kwargs,
+                num_proc=num_proc,
             )
 
         results = {qid: {} for qid in query_idx_to_id.values()}
@@ -144,6 +150,7 @@ class PylateSearchEncoder:
         hf_split: str,
         top_k: int,
         encode_kwargs: EncodeKwargs,
+        num_proc: int,
     ) -> dict[str, list[tuple[float, str]]]:
         from pylate import indexes, retrieve
 
@@ -170,6 +177,7 @@ class PylateSearchEncoder:
             task_metadata,
             prompt_type=PromptType.document,
             batch_size=encode_kwargs.get("batch_size", 32),
+            num_proc=num_proc,
         )
         documents_embeddings = self.encode(
             documents_loader,
@@ -208,6 +216,7 @@ class PylateSearchEncoder:
         hf_subset: str,
         hf_split: str,
         encode_kwargs: EncodeKwargs,
+        num_proc: int = 1,
     ) -> dict[str, list[tuple[float, str]]]:
         """Rerank with PyLate's rank.rerank using per-query candidates.
 
@@ -230,6 +239,7 @@ class PylateSearchEncoder:
                 task_metadata,
                 prompt_type=PromptType.document,
                 batch_size=encode_kwargs.get("batch_size", 32),
+                num_proc=num_proc,
             ),
             task_metadata=task_metadata,
             hf_split=hf_split,
@@ -352,6 +362,7 @@ colbert_v2 = ModelMeta(
     public_training_data=None,
     release_date="2024-09-21",
     n_parameters=int(110 * 1e6),
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=418,
     max_tokens=180,
     embed_dim=None,
@@ -408,6 +419,7 @@ jina_colbert_v2 = ModelMeta(
     public_training_data=None,
     release_date="2024-08-16",
     n_parameters=int(559 * 1e6),
+    n_embedding_parameters=None,
     memory_usage_mb=1067,
     max_tokens=8192,
     embed_dim=None,
@@ -464,6 +476,7 @@ lightonai__gte_moderncolbert_v1 = ModelMeta(
     public_training_data="https://huggingface.co/datasets/lightonai/ms-marco-en-bge-gemma",
     release_date="2025-04-30",
     n_parameters=int(149 * 1e6),
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=8192,
     embed_dim=None,

mteb/models/model_implementations/qodo_models.py

@@ -36,6 +36,7 @@ Qodo_Embed_1_1_5B = ModelMeta(
     revision="84bbef079b32e8823ec226d4e9e92902706b0eb6",
     release_date="2025-02-19",
     n_parameters=1_780_000_000,
+    n_embedding_parameters=232_928_256,
     memory_usage_mb=6776,
     embed_dim=1536,
     license="https://huggingface.co/Qodo/Qodo-Embed-1-1.5B/blob/main/LICENSE",
@@ -59,6 +60,7 @@ Qodo_Embed_1_7B = ModelMeta(
     revision="f9edd9bf7f687c0e832424058e265120f603cd81",
     release_date="2025-02-24",
     n_parameters=7_613_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=29040,
     embed_dim=3584,
     license="https://huggingface.co/Qodo/Qodo-Embed-1-1.5B/blob/main/LICENSE",

mteb/models/model_implementations/qtack_models.py

@@ -31,6 +31,7 @@ mini_gte = ModelMeta(
     revision="7fbe6f9b4cc42615e0747299f837ad7769025492",
     release_date="2025-01-28",
     n_parameters=int(66.3 * 1e6),
+    n_embedding_parameters=23_440_896,
     memory_usage_mb=253,
     embed_dim=768,
     license="apache-2.0",

mteb/models/model_implementations/qwen3_models.py

@@ -147,6 +147,7 @@ Qwen3_Embedding_0B6 = ModelMeta(
     revision="b22da495047858cce924d27d76261e96be6febc0",  # Commit of @tomaarsen
     release_date="2025-06-05",
     n_parameters=595776512,
+    n_embedding_parameters=None,
     memory_usage_mb=1136,
     embed_dim=1024,
     max_tokens=32768,
@@ -170,6 +171,7 @@ Qwen3_Embedding_4B = ModelMeta(
     revision="636cd9bf47d976946cdbb2b0c3ca0cb2f8eea5ff",  # Commit of @tomaarsen
     release_date="2025-06-05",
     n_parameters=4021774336,
+    n_embedding_parameters=None,
     memory_usage_mb=7671,
     embed_dim=2560,
     max_tokens=32768,
@@ -193,6 +195,7 @@ Qwen3_Embedding_8B = ModelMeta(
     revision="4e423935c619ae4df87b646a3ce949610c66241c",  # Commit of @tomaarsen
     release_date="2025-06-05",
     n_parameters=7567295488,
+    n_embedding_parameters=None,
     memory_usage_mb=14433,
     embed_dim=4096,
     max_tokens=32768,

mteb/models/model_implementations/qzhou_models.py

@@ -64,6 +64,7 @@ QZhou_Embedding = ModelMeta(
     revision="f1e6c03ee3882e7b9fa5cec91217715272e433b8",
     release_date="2025-08-24",
     n_parameters=7_070_619_136,
+    n_embedding_parameters=None,
     memory_usage_mb=14436,
     embed_dim=3584,
     license="apache-2.0",
@@ -98,6 +99,7 @@ QZhou_Embedding_Zh = ModelMeta(
     revision="0321ccb126413d1e49c5ce908e802b63d35f18e2",
     release_date="2025-09-28",
     n_parameters=7_575_747_328,
+    n_embedding_parameters=None,
     memory_usage_mb=29431,
     embed_dim=1792,
     license="apache-2.0",

mteb/models/model_implementations/rasgaard_models.py

@@ -12,6 +12,7 @@ potion_base_8m = ModelMeta(
     revision="387897cfb09992e6d45ea9cd7b28b9fcf119e23a",
     release_date="2025-10-08",
     n_parameters=22893312,
+    n_embedding_parameters=22893312,
     memory_usage_mb=87,
     max_tokens=np.inf,
     embed_dim=256,