mteb 2.7.2__py3-none-any.whl → 2.7.4__py3-none-any.whl

mteb/models/model_implementations/kowshik24_models.py

@@ -9,6 +9,7 @@ kowshik24_bangla_embedding_model = ModelMeta(
     revision="6689c21e69be5950596bad084457cbaa138728d8",
     release_date="2025-11-10",
     n_parameters=278_000_000,
+    n_embedding_parameters=192_001_536,
     memory_usage_mb=1061,
     embed_dim=768,
     license="apache-2.0",

mteb/models/model_implementations/lens_models.py

@@ -18,6 +18,7 @@ lens_d4000 = ModelMeta(
     revision="e473b33364e6c48a324796fd1411d3b93670c6fe",
     release_date="2025-01-17",
     n_parameters=int(7.11 * 1e9),
+    n_embedding_parameters=131_084_288,
     memory_usage_mb=27125,
     embed_dim=4000,
     license="apache-2.0",
@@ -41,6 +42,7 @@ lens_d8000 = ModelMeta(
     revision="a0b87bd91cb27b6f2f0b0fe22c28026da1d464ef",
     release_date="2025-01-17",
     n_parameters=int(7.11 * 1e9),
+    n_embedding_parameters=131_084_288,
     memory_usage_mb=27125,
     embed_dim=8000,
     license="apache-2.0",

mteb/models/model_implementations/lgai_embedding_models.py

@@ -52,6 +52,7 @@ lgai_embedding_en = ModelMeta(
     revision="5e0b2316acc8c2e2941ded6b9cb200b1cb313e65",
     release_date="2025-06-11",
     n_parameters=7_110_000_000,
+    n_embedding_parameters=131_084_288,
     memory_usage_mb=27125,
     embed_dim=4096,
     license="apache-2.0",

mteb/models/model_implementations/linq_models.py

@@ -1,11 +1,16 @@
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
 import torch
 
 from mteb.models.instruct_wrapper import instruct_wrapper
 from mteb.models.model_meta import ModelMeta, ScoringFunction
-from mteb.types import PromptType
 
 from .e5_instruct import E5_MISTRAL_TRAINING_DATA
 
+if TYPE_CHECKING:
+    from mteb.types import PromptType
 LINQ_EMBED_MISTRAL_CITATION = """@misc{LinqAIResearch2024,
   title={Linq-Embed-Mistral:Elevating Text Retrieval with Improved GPT Data Through Task-Specific Control and Quality Refinement},
   author={Junseong Kim and Seolhwa Lee and Jihoon Kwon and Sangmo Gu and Yejin Kim and Minkyung Cho and Jy-yong Sohn and Chanyeol Choi},
@@ -38,6 +43,7 @@ Linq_Embed_Mistral = ModelMeta(
     revision="0c1a0b0589177079acc552433cad51d7c9132379",
     release_date="2024-05-29",  # initial commit of hf model.
     n_parameters=7_110_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=13563,
     embed_dim=4096,
     license="cc-by-nc-4.0",

mteb/models/model_implementations/listconranker.py

@@ -1,14 +1,19 @@
-from typing import Any
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any
 
 import torch
-from torch.utils.data import DataLoader
 
-from mteb.abstasks.task_metadata import TaskMetadata
 from mteb.models.model_meta import ModelMeta
-from mteb.types import BatchedInput, PromptType
 
 from .rerankers_custom import RerankerWrapper
 
+if TYPE_CHECKING:
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import BatchedInput, PromptType
+
 LISTCONRANKER_CITATION = """@article{liu2025listconranker,
   title={ListConRanker: A Contrastive Text Reranker with Listwise Encoding},
   author={Liu, Junlong and Ma, Yue and Zhao, Ruihui and Zheng, Junhao and Ma, Qianli and Kang, Yangyang},
@@ -118,6 +123,7 @@ listconranker = ModelMeta(
     revision="95ae6a5f422a916bc36520f0f3e198e7d91520a0",
     release_date="2024-12-11",
     n_parameters=401_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=1242,
     similarity_fn_name="cosine",
     training_datasets=listconranker_training_datasets,

mteb/models/model_implementations/llm2clip_models.py

@@ -1,15 +1,20 @@
+from __future__ import annotations
+
 from pathlib import Path
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import torch
-from torch.utils.data import DataLoader
 from tqdm.auto import tqdm
 
 from mteb._requires_package import requires_image_dependencies, requires_package
-from mteb.abstasks.task_metadata import TaskMetadata
 from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta, ScoringFunction
-from mteb.types import Array, BatchedInput, PromptType
+
+if TYPE_CHECKING:
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput, PromptType
 
 LLM2CLIP_CITATION = """@misc{huang2024llm2clippowerfullanguagemodel,
   title={LLM2CLIP: Powerful Language Model Unlock Richer Visual Representation},
@@ -189,6 +194,7 @@ llm2clip_openai_l_14_336 = ModelMeta(
     release_date="2024-11-07",
     modalities=["image", "text"],
     n_parameters=579_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=None,
     embed_dim=1280,
@@ -214,6 +220,7 @@ llm2clip_openai_l_14_224 = ModelMeta(
     release_date="2024-11-07",
     modalities=["image", "text"],
     n_parameters=578_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=None,
     embed_dim=1280,
@@ -238,6 +245,7 @@ llm2clip_openai_b_16 = ModelMeta(
     release_date="2024-11-07",
     modalities=["image", "text"],
     n_parameters=361_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=None,
     max_tokens=None,
     embed_dim=1280,

mteb/models/model_implementations/llm2vec_models.py

@@ -1,16 +1,22 @@
+from __future__ import annotations
+
 import logging
-from collections.abc import Callable
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import torch
-from torch.utils.data import DataLoader
 
 from mteb._requires_package import requires_package, suggest_package
-from mteb.abstasks.task_metadata import TaskMetadata
 from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta, ScoringFunction
-from mteb.models.models_protocols import EncoderProtocol
-from mteb.types import Array, BatchedInput, PromptType
+
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.models.models_protocols import EncoderProtocol
+    from mteb.types import Array, BatchedInput, PromptType
 
 logger = logging.getLogger(__name__)
 
@@ -139,6 +145,7 @@ llm2vec_llama3_8b_supervised = ModelMeta(
     # TODO: Not sure what to put here as a model is made of two peft repos, each with a different revision
     release_date="2024-04-09",
     n_parameters=7_505_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=28629,
     max_tokens=8192,
     embed_dim=4096,
@@ -168,6 +175,7 @@ llm2vec_llama3_8b_unsupervised = ModelMeta(
     revision="1cb7b735326d13a8541db8f57f35da5373f5e9c6",
     release_date="2024-04-09",
     n_parameters=7_505_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=28629,
     max_tokens=8192,
     embed_dim=4096,
@@ -197,6 +205,7 @@ llm2vec_mistral7b_supervised = ModelMeta(
     revision="0ae69bdd5816105778b971c3138e8f8a18eaa3ae",
     release_date="2024-04-09",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=131_072_000,
     memory_usage_mb=27126,
     max_tokens=32768,
     embed_dim=4096,
@@ -226,6 +235,7 @@ llm2vec_mistral7b_unsupervised = ModelMeta(
     revision="2c055a5d77126c0d3dc6cd8ffa30e2908f4f45f8",
     release_date="2024-04-09",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=131_072_000,
     memory_usage_mb=27126,
     max_tokens=32768,
     embed_dim=4096,
@@ -255,6 +265,7 @@ llm2vec_llama2_7b_supervised = ModelMeta(
     revision="2c055a5d77126c0d3dc6cd8ffa30e2908f4f45f8",
     release_date="2024-04-09",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=27126,
     max_tokens=32768,
     embed_dim=4096,
@@ -284,6 +295,7 @@ llm2vec_llama2_7b_unsupervised = ModelMeta(
     revision="a76944871d169ebe7c97eb921764cd063afed785",
     release_date="2024-04-09",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=None,
     memory_usage_mb=27126,
     max_tokens=32768,
     embed_dim=4096,
@@ -313,6 +325,7 @@ llm2vec_sheared_llama_supervised = ModelMeta(
     revision="a5943d406c6b016fef3f07906aac183cf1a0b47d",
     release_date="2024-04-09",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=65_536_000,
     memory_usage_mb=27126,
     max_tokens=32768,
     embed_dim=4096,
@@ -342,6 +355,7 @@ llm2vec_sheared_llama_unsupervised = ModelMeta(
     revision="a5943d406c6b016fef3f07906aac183cf1a0b47d",
     release_date="2024-04-09",
     n_parameters=7_111_000_000,
+    n_embedding_parameters=65_536_000,
     memory_usage_mb=27126,
     max_tokens=32768,
     embed_dim=4096,

mteb/models/model_implementations/mcinext_models.py

@@ -1,16 +1,19 @@
+from __future__ import annotations
+
 import logging
 import os
 import time
 import warnings
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import numpy as np
 import requests
 
 from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta
-from mteb.types import PromptType
 
+if TYPE_CHECKING:
+    from mteb.types import PromptType
 logger = logging.getLogger(__name__)
 
 HAKIM_CITATION = """@article{sarmadi2025hakim,
@@ -355,6 +358,7 @@ hakim = ModelMeta(
     revision="1",
     release_date="2025-05-10",
     n_parameters=124_441_344,
+    n_embedding_parameters=None,
     memory_usage_mb=475,
     embed_dim=768,
     license="not specified",
@@ -423,6 +427,7 @@ hakim_small = ModelMeta(
     revision="1",
     release_date="2025-05-10",
     n_parameters=38_736_384,
+    n_embedding_parameters=None,
     memory_usage_mb=148,
     embed_dim=512,
     license="not specified",
@@ -490,6 +495,7 @@ hakim_unsup = ModelMeta(
     revision="1",
     release_date="2025-05-10",
     n_parameters=124_441_344,
+    n_embedding_parameters=None,
     memory_usage_mb=475,
     embed_dim=768,
     license="not specified",

mteb/models/model_implementations/mdbr_models.py

@@ -45,6 +45,7 @@ mdbr_leaf_ir = ModelMeta(
         "Transformers",
     ],
     n_parameters=22_861_056,
+    n_embedding_parameters=11_720_448,
     memory_usage_mb=86,
     max_tokens=512,
     embed_dim=768,
@@ -79,6 +80,7 @@ mdbr_leaf_mt = ModelMeta(
         "Transformers",
     ],
     n_parameters=22_958_592,
+    n_embedding_parameters=11_720_448,
     memory_usage_mb=86,
     max_tokens=512,
     embed_dim=1024,