mteb 2.7.1__py3-none-any.whl → 2.7.3__py3-none-any.whl

mteb/models/abs_encoder.py

@@ -1,14 +1,12 @@
+from __future__ import annotations
+
 import logging
 import warnings
 from abc import ABC, abstractmethod
-from collections.abc import Callable, Sequence
-from typing import Any, Literal, cast, get_args, overload
-
-from torch.utils.data import DataLoader
-from typing_extensions import Unpack
+from typing import TYPE_CHECKING, Any, Literal, cast, get_args, overload
 
 import mteb
-from mteb.abstasks.task_metadata import TaskMetadata, TaskType
+from mteb.abstasks.task_metadata import TaskType
 from mteb.similarity_functions import (
     cos_sim,
     dot_score,
@@ -18,13 +16,25 @@ from mteb.similarity_functions import (
     pairwise_max_sim,
 )
 from mteb.types import (
-    Array,
-    BatchedInput,
-    EncodeKwargs,
     PromptType,
 )
 
-from .model_meta import ModelMeta, ScoringFunction
+from .model_meta import ScoringFunction
+
+if TYPE_CHECKING:
+    from collections.abc import Callable, Sequence
+
+    from torch.utils.data import DataLoader
+    from typing_extensions import Unpack
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import (
+        Array,
+        BatchedInput,
+        EncodeKwargs,
+    )
+
+    from .model_meta import ModelMeta
 
 logger = logging.getLogger(__name__)
 
@@ -314,7 +324,7 @@ class AbsEncoder(ABC):
             ):
                 arr = self.model.similarity(embeddings1, embeddings2)
                 # We assume that the model returns an Array-like object:
-                arr = cast(Array, arr)
+                arr = cast("Array", arr)
                 return arr
             return cos_sim(embeddings1, embeddings2)
         if self.mteb_model_meta.similarity_fn_name is ScoringFunction.COSINE:
@@ -352,7 +362,7 @@ class AbsEncoder(ABC):
             ):
                 arr = self.model.similarity_pairwise(embeddings1, embeddings2)
                 # We assume that the model returns an Array-like object:
-                arr = cast(Array, arr)
+                arr = cast("Array", arr)
                 return arr
             return pairwise_cos_sim(embeddings1, embeddings2)
         if self.mteb_model_meta.similarity_fn_name is ScoringFunction.COSINE:

mteb/models/cache_wrappers/cache_backend_protocol.py

@@ -1,9 +1,11 @@
 from __future__ import annotations
 
-from pathlib import Path
-from typing import Any, Protocol, runtime_checkable
+from typing import TYPE_CHECKING, Any, Protocol, runtime_checkable
 
-import numpy as np
+if TYPE_CHECKING:
+    from pathlib import Path
+
+    import numpy as np
 
 
 @runtime_checkable

mteb/models/cache_wrappers/cache_backends/_hash_utils.py

@@ -1,6 +1,12 @@
+from __future__ import annotations
+
 import hashlib
-from collections.abc import Mapping
-from typing import Any
+from typing import TYPE_CHECKING, Any
+
+if TYPE_CHECKING:
+    from collections.abc import Mapping
+
+    from PIL import Image
 
 
 def _hash_item(item: Mapping[str, Any]) -> str:
@@ -10,8 +16,6 @@ def _hash_item(item: Mapping[str, Any]) -> str:
         item_hash = hashlib.sha256(item_text.encode()).hexdigest()
 
     if "image" in item:
-        from PIL import Image
-
         image: Image.Image = item["image"]
         item_hash += hashlib.sha256(image.tobytes()).hexdigest()
 

mteb/models/cache_wrappers/cache_backends/faiss_cache.py

@@ -1,16 +1,22 @@
+from __future__ import annotations
+
 import json
 import logging
 import warnings
 from pathlib import Path
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import numpy as np
 
 from mteb._requires_package import requires_package
-from mteb.types import BatchedInput
 
 from ._hash_utils import _hash_item
 
+if TYPE_CHECKING:
+    import faiss
+
+    from mteb.types import BatchedInput
+
 logger = logging.getLogger(__name__)
 
 
@@ -24,7 +30,6 @@ class FaissCache:
             "FAISS-based vector cache",
             install_instruction="pip install mteb[faiss-cpu]",
         )
-        import faiss
 
         self.directory = Path(directory)
         self.directory.mkdir(parents=True, exist_ok=True)

mteb/models/cache_wrappers/cache_wrapper.py

@@ -1,21 +1,26 @@
+from __future__ import annotations
+
 import logging
 from pathlib import Path
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import numpy as np
 import torch
 from datasets import Dataset
-from torch.utils.data import DataLoader
 
 from mteb._create_dataloaders import create_dataloader
-from mteb.abstasks.task_metadata import TaskMetadata
-from mteb.models.cache_wrappers.cache_backend_protocol import (
-    CacheBackendProtocol,
-)
 from mteb.models.cache_wrappers.cache_backends.numpy_cache import NumpyCache
-from mteb.models.model_meta import ModelMeta
-from mteb.models.models_protocols import EncoderProtocol
-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.models.cache_wrappers.cache_backend_protocol import (
+        CacheBackendProtocol,
+    )
+    from mteb.models.model_meta import ModelMeta
+    from mteb.models.models_protocols import EncoderProtocol
+    from mteb.types import Array, BatchedInput, PromptType
 
 logger = logging.getLogger(__name__)
 

mteb/models/get_model_meta.py

@@ -1,15 +1,22 @@
+from __future__ import annotations
+
 import difflib
 import logging
-from collections.abc import Iterable
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
-from mteb.abstasks import AbsTask
 from mteb.models import (
     ModelMeta,
-    MTEBModels,
 )
 from mteb.models.model_implementations import MODEL_REGISTRY
 
+if TYPE_CHECKING:
+    from collections.abc import Iterable
+
+    from mteb.abstasks import AbsTask
+    from mteb.models import (
+        MTEBModels,
+    )
+
 logger = logging.getLogger(__name__)
 
 

mteb/models/instruct_wrapper.py

@@ -1,16 +1,24 @@
+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
-from mteb.abstasks.task_metadata import TaskMetadata
-from mteb.types import Array, BatchedInput, PromptType
+from mteb.types import PromptType
 
 from .abs_encoder import AbsEncoder
 
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput
+
+
 logger = logging.getLogger(__name__)
 
 

mteb/models/model_implementations/align_models.py

@@ -1,13 +1,18 @@
-from typing import Any
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any
 
 import torch
-from torch.utils.data import DataLoader
 from tqdm.auto import tqdm
 
-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
 
 
 class ALIGNModel(AbsEncoder):

mteb/models/model_implementations/bedrock_models.py

@@ -1,20 +1,30 @@
+from __future__ import annotations
+
 import json
 import logging
 import re
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import numpy as np
-from torch.utils.data import DataLoader
 from tqdm.auto import tqdm
 
 from mteb._requires_package import 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
 
-from .cohere_models import model_prompts as cohere_model_prompts
-from .cohere_models import supported_languages as cohere_supported_languages
+from .cohere_models import (
+    model_prompts as cohere_model_prompts,
+)
+from .cohere_models import (
+    supported_languages as cohere_supported_languages,
+)
+
+if TYPE_CHECKING:
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput, PromptType
+
 
 logger = logging.getLogger(__name__)
 

mteb/models/model_implementations/blip2_models.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 tqdm.auto import tqdm
 
 from mteb._requires_package import 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
 
 BLIP2_CITATION = """@inproceedings{li2023blip2,
     title={{BLIP-2:} Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models},

mteb/models/model_implementations/blip_models.py

@@ -1,14 +1,19 @@
-from typing import Any
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any
 
 import torch
 from torch.nn.functional import normalize
-from torch.utils.data import DataLoader
 from tqdm.auto import tqdm
 
-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
 
 BLIP_CITATION = """@misc{https://doi.org/10.48550/arxiv.2201.12086,
     doi = {10.48550/ARXIV.2201.12086},

mteb/models/model_implementations/bm25.py

@@ -1,18 +1,23 @@
+from __future__ import annotations
+
 import logging
+from typing import TYPE_CHECKING
 
 from mteb._create_dataloaders import _create_text_queries_dataloader
 from mteb._requires_package import requires_package
-from mteb.abstasks.task_metadata import TaskMetadata
 from mteb.models.model_meta import ModelMeta
-from mteb.models.models_protocols import SearchProtocol
-from mteb.types import (
-    CorpusDatasetType,
-    EncodeKwargs,
-    InstructionDatasetType,
-    QueryDatasetType,
-    RetrievalOutputType,
-    TopRankedDocumentsType,
-)
+
+if TYPE_CHECKING:
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.models.models_protocols import SearchProtocol
+    from mteb.types import (
+        CorpusDatasetType,
+        EncodeKwargs,
+        InstructionDatasetType,
+        QueryDatasetType,
+        RetrievalOutputType,
+        TopRankedDocumentsType,
+    )
 
 logger = logging.getLogger(__name__)
 

mteb/models/model_implementations/bmretriever_models.py

@@ -1,5 +1,6 @@
-from collections.abc import Callable
-from typing import Any
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any
 
 import torch
 from sentence_transformers import SentenceTransformer
@@ -9,6 +10,9 @@ from mteb.models import ModelMeta
 from mteb.models.instruct_wrapper import InstructSentenceTransformerModel
 from mteb.types import PromptType
 
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
 
 def instruction_template(
     instruction: str, prompt_type: PromptType | None = None

mteb/models/model_implementations/cde_models.py

@@ -1,27 +1,31 @@
+from __future__ import annotations
+
 import logging
-from collections.abc import Sequence
 from typing import TYPE_CHECKING, Any
 
 import numpy as np
 import torch
-from torch.utils.data import DataLoader
 
 import mteb
 from mteb._create_dataloaders import _corpus_to_dict
-from mteb.abstasks.task_metadata import TaskMetadata
 from mteb.models.model_meta import ModelMeta, ScoringFunction
-from mteb.models.models_protocols import PromptType
 from mteb.models.sentence_transformer_wrapper import SentenceTransformerEncoderWrapper
-from mteb.types import Array, BatchedInput
+from mteb.types import PromptType
 
 from .bge_models import bge_full_data
 
 if TYPE_CHECKING:
+    from collections.abc import Sequence
+
+    from torch.utils.data import DataLoader
+
     from mteb.abstasks import (
         AbsTaskClassification,
         AbsTaskRetrieval,
         AbsTaskSummarization,
     )
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput
 logger = logging.getLogger(__name__)
 
 CDE_CITATION = """@misc{morris2024contextualdocumentembeddings,

mteb/models/model_implementations/clip_models.py

@@ -1,13 +1,18 @@
-from typing import Any
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any
 
 import torch
-from torch.utils.data import DataLoader
 from tqdm.auto import tqdm
 
-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
 
 
 class CLIPModel(AbsEncoder):

mteb/models/model_implementations/cohere_models.py

@@ -1,18 +1,24 @@
+from __future__ import annotations
+
 import logging
 import time
 from functools import wraps
-from typing import Any, Literal, get_args
+from typing import TYPE_CHECKING, Any, Literal, get_args
 
 import numpy as np
 import torch
-from torch.utils.data import DataLoader
 from tqdm.auto import tqdm
 
 from mteb._requires_package import 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
+from mteb.types import PromptType
+
+if TYPE_CHECKING:
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput
 
 logger = logging.getLogger(__name__)
 

mteb/models/model_implementations/cohere_v.py

@@ -1,15 +1,15 @@
+from __future__ import annotations
+
 import base64
 import io
 import os
 import time
-from typing import Any, Literal, get_args
+from typing import TYPE_CHECKING, Any, Literal, get_args
 
 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 import ModelMeta
 from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_implementations.cohere_models import (
@@ -18,7 +18,12 @@ from mteb.models.model_implementations.cohere_models import (
     retry_with_rate_limit,
 )
 from mteb.models.model_meta import 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
 
 
 def _post_process_embeddings(

mteb/models/model_implementations/colpali_models.py

@@ -4,20 +4,21 @@ import logging
 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 PIL import Image
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput, PromptType
 
 logger = logging.getLogger(__name__)
 

mteb/models/model_implementations/colqwen_models.py

@@ -1,18 +1,23 @@
+from __future__ import annotations
+
 import logging
-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
 
 from .colpali_models import (
     COLPALI_CITATION,
@@ -329,32 +334,6 @@ colqwen3_4b = ModelMeta(
     citation=TOMORO_CITATION,
 )
 
-colnomic_7b = ModelMeta(
-    loader=ColQwen2_5Wrapper,
-    loader_kwargs=dict(
-        torch_dtype=torch.float16,
-    ),
-    name="nomic-ai/colnomic-embed-multimodal-7b",
-    model_type=["late-interaction"],
-    languages=["eng-Latn"],
-    revision="530094e83a40ca4edcb5c9e5ddfa61a4b5ea0d2f",
-    release_date="2025-03-31",
-    modalities=["image", "text"],
-    n_parameters=7_000_000_000,
-    memory_usage_mb=14400,
-    max_tokens=128000,
-    embed_dim=128,
-    license="apache-2.0",
-    open_weights=True,
-    public_training_code="https://github.com/nomic-ai/colpali",
-    public_training_data="https://huggingface.co/datasets/vidore/colpali_train_set",
-    framework=["ColPali", "safetensors"],
-    reference="https://huggingface.co/nomic-ai/colnomic-embed-multimodal-7b",
-    similarity_fn_name="MaxSim",
-    use_instructions=True,
-    training_datasets=COLPALI_TRAINING_DATA,
-    citation=COLPALI_CITATION,
-)
 
 COLNOMIC_CITATION = """
 @misc{nomicembedmultimodal2025,
@@ -402,7 +381,7 @@ colnomic_3b = ModelMeta(
 )
 
 colnomic_7b = ModelMeta(
-    loader=ColQwen2Wrapper,
+    loader=ColQwen2_5Wrapper,
     loader_kwargs=dict(
         torch_dtype=torch.float16,
     ),

mteb/models/model_implementations/colsmol_models.py

@@ -56,7 +56,7 @@ colsmol_256m = ModelMeta(
     name="vidore/colSmol-256M",
     model_type=["late-interaction"],
     languages=["eng-Latn"],
-    revision="530094e83a40ca4edcb5c9e5ddfa61a4b5ea0d2f",
+    revision="a59110fdf114638b8018e6c9a018907e12f14855",
     release_date="2025-01-22",
     modalities=["image", "text"],
     n_parameters=256_000_000,

mteb/models/model_implementations/conan_models.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import hashlib
 import json
 import logging
@@ -5,20 +7,24 @@ import os
 import random
 import string
 import time
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import numpy as np
 import requests
-from torch.utils.data import DataLoader
 
-from mteb.abstasks.task_metadata import TaskMetadata
 from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta
-from mteb.types import Array, BatchedInput, PromptType
 
 from .bge_models import bge_full_data
 from .e5_instruct import E5_MISTRAL_TRAINING_DATA
 
+if TYPE_CHECKING:
+    from torch.utils.data import DataLoader
+
+    from mteb.abstasks.task_metadata import TaskMetadata
+    from mteb.types import Array, BatchedInput, PromptType
+
+
 conan_zh_datasets = {
     "BQ",
     "LCQMC",