mteb 2.5.2__py3-none-any.whl → 2.7.2__py3-none-any.whl

mteb/models/model_implementations/bge_models.py

@@ -331,7 +331,13 @@ bge_small_en_v1_5 = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-small-en-v1.5",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=[
+        "Sentence Transformers",
+        "PyTorch",
+        "ONNX",
+        "safetensors",
+        "Transformers",
+    ],
     use_instructions=True,
     public_training_code=None,
     public_training_data="https://data.baai.ac.cn/details/BAAI-MTP",
@@ -357,7 +363,13 @@ bge_base_en_v1_5 = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-base-en-v1.5",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=[
+        "Sentence Transformers",
+        "PyTorch",
+        "ONNX",
+        "safetensors",
+        "Transformers",
+    ],
     use_instructions=True,
     public_training_code=None,  # seemingly released (at least for some models, but the link is broken
     public_training_data="https://data.baai.ac.cn/details/BAAI-MTP",
@@ -383,7 +395,13 @@ bge_large_en_v1_5 = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-large-en-v1.5",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=[
+        "Sentence Transformers",
+        "PyTorch",
+        "ONNX",
+        "safetensors",
+        "Transformers",
+    ],
     use_instructions=True,
     citation=BGE_15_CITATION,
     public_training_code=None,  # seemingly released (at least for some models, but the link is broken
@@ -409,7 +427,7 @@ bge_small_zh = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-small-zh",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -436,7 +454,7 @@ bge_base_zh = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-base-zh",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -463,7 +481,7 @@ bge_large_zh = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-large-zh",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -490,7 +508,7 @@ bge_small_en = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-small-en",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data="https://data.baai.ac.cn/details/BAAI-MTP",
@@ -517,7 +535,13 @@ bge_base_en = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-base-en",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=[
+        "Sentence Transformers",
+        "PyTorch",
+        "Transformers",
+        "ONNX",
+        "safetensors",
+    ],
     use_instructions=True,
     public_training_code=None,  # seemingly released (at least for some models, but the link is broken
     public_training_data="https://data.baai.ac.cn/details/BAAI-MTP",
@@ -544,7 +568,7 @@ bge_large_en = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-large-en",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,  # seemingly released (at least for some models, but the link is broken
     public_training_data="https://data.baai.ac.cn/details/BAAI-MTP",
@@ -572,7 +596,7 @@ bge_small_zh_v1_5 = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-small-zh-v1.5",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -598,7 +622,7 @@ bge_base_zh_v1_5 = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-base-zh-v1.5",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -624,7 +648,7 @@ bge_large_zh_v1_5 = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/BAAI/bge-large-zh-v1.5",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -647,13 +671,13 @@ bge_m3 = ModelMeta(
     max_tokens=8194,
     reference="https://huggingface.co/BAAI/bge-m3",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "ONNX"],
     use_instructions=False,
     public_training_code=None,
     public_training_data="https://huggingface.co/datasets/cfli/bge-full-data",
     training_datasets=bge_m3_training_data,
     citation="""@misc{bge-m3,
-      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation}, 
+      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation},
       author={Jianlv Chen and Shitao Xiao and Peitian Zhang and Kun Luo and Defu Lian and Zheng Liu},
       year={2024},
       eprint={2402.03216},
@@ -743,7 +767,7 @@ bge_multilingual_gemma2 = ModelMeta(
     max_tokens=8192,  # from old C-MTEB leaderboard
     reference="https://huggingface.co/BAAI/bge-multilingual-gemma2",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=False,
     public_training_code=None,
     public_training_data=None,
@@ -754,7 +778,7 @@ bge_multilingual_gemma2 = ModelMeta(
     | bge_full_data
     | bge_m3_training_data,
     citation="""@misc{bge-m3,
-      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation}, 
+      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation},
       author={Jianlv Chen and Shitao Xiao and Peitian Zhang and Kun Luo and Defu Lian and Zheng Liu},
       year={2024},
       eprint={2402.03216},
@@ -764,7 +788,7 @@ bge_multilingual_gemma2 = ModelMeta(
 
 
 @misc{bge_embedding,
-      title={C-Pack: Packaged Resources To Advance General Chinese Embedding}, 
+      title={C-Pack: Packaged Resources To Advance General Chinese Embedding},
       author={Shitao Xiao and Zheng Liu and Peitian Zhang and Niklas Muennighoff},
       year={2023},
       eprint={2309.07597},
@@ -790,7 +814,7 @@ bge_en_icl = ModelMeta(
     max_tokens=32768,
     reference="https://huggingface.co/BAAI/bge-en-icl",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=False,
     public_training_code="https://github.com/FlagOpen/FlagEmbedding",
     public_training_data="https://huggingface.co/datasets/cfli/bge-full-data",
@@ -824,13 +848,13 @@ bge_m3_unsupervised = ModelMeta(
     max_tokens=8192,
     reference="https://huggingface.co/BAAI/bge-m3-unsupervised",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors"],
     use_instructions=False,
     public_training_code="https://github.com/FlagOpen/FlagEmbedding",
     public_training_data="https://huggingface.co/datasets/cfli/bge-full-data",
     training_datasets=bge_m3_training_data,
     citation="""@misc{bge-m3,
-      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation}, 
+      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation},
       author={Jianlv Chen and Shitao Xiao and Peitian Zhang and Kun Luo and Defu Lian and Zheng Liu},
       year={2024},
       eprint={2402.03216},
@@ -854,7 +878,7 @@ manu__bge_m3_custom_fr = ModelMeta(
     open_weights=True,
     public_training_code=None,
     public_training_data=None,
-    framework=["PyTorch", "Sentence Transformers"],
+    framework=["PyTorch", "Sentence Transformers", "safetensors"],
     reference="https://huggingface.co/manu/bge-m3-custom-fr",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=None,

mteb/models/model_implementations/bica_model.py

@@ -15,20 +15,20 @@ bica_base = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/bisectgroup/BiCA-base",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors"],
     use_instructions=False,
     public_training_code="https://github.com/NiravBhattLab/BiCA",
     public_training_data="https://huggingface.co/datasets/bisectgroup/hard-negatives-traversal",
     adapted_from="thenlper/gte-base",
     citation="""
 @misc{sinha2025bicaeffectivebiomedicaldense,
-      title={BiCA: Effective Biomedical Dense Retrieval with Citation-Aware Hard Negatives}, 
+      title={BiCA: Effective Biomedical Dense Retrieval with Citation-Aware Hard Negatives},
       author={Aarush Sinha and Pavan Kumar S and Roshan Balaji and Nirav Pravinbhai Bhatt},
       year={2025},
       eprint={2511.08029},
       archivePrefix={arXiv},
       primaryClass={cs.IR},
-      url={https://arxiv.org/abs/2511.08029}, 
+      url={https://arxiv.org/abs/2511.08029},
 }
 """,
     training_datasets=set(),

mteb/models/model_implementations/blip2_models.py

@@ -179,7 +179,7 @@ blip2_opt_2_7b = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/LAVIS/tree/main/projects/blip2",
     public_training_data=None,
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers", "safetensors"],
     reference="https://huggingface.co/Salesforce/blip2-opt-2.7b",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -203,7 +203,7 @@ blip2_opt_6_7b_coco = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/LAVIS/tree/main/projects/blip2",
     public_training_data=None,
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers", "safetensors"],
     reference="https://huggingface.co/Salesforce/blip2-opt-6.7b-coco",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,

mteb/models/model_implementations/blip_models.py

@@ -128,7 +128,7 @@ class BLIPModel(AbsEncoder):
 
 # in descending order of usage (downloads from huggingface)
 blip_image_captioning_large = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-image-captioning-large",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -143,7 +143,7 @@ blip_image_captioning_large = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers", "safetensors"],
     reference="https://huggingface.co/Salesforce/blip-image-captioning-large",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -156,7 +156,7 @@ blip_image_captioning_large = ModelMeta(
 )
 
 blip_image_captioning_base = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-image-captioning-base",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -171,7 +171,7 @@ blip_image_captioning_base = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/Salesforce/blip-image-captioning-base",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -185,7 +185,7 @@ blip_image_captioning_base = ModelMeta(
 
 
 blip_vqa_base = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-vqa-base",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -200,7 +200,7 @@ blip_vqa_base = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers", "safetensors"],
     reference="https://huggingface.co/Salesforce/blip-vqa-base",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -212,7 +212,7 @@ blip_vqa_base = ModelMeta(
 )
 
 blip_vqa_capfilt_large = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-vqa-capfilt-large",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -227,7 +227,7 @@ blip_vqa_capfilt_large = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/Salesforce/blip-vqa-capfilt-large",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -239,7 +239,7 @@ blip_vqa_capfilt_large = ModelMeta(
 )
 
 blip_itm_base_coco = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-itm-base-coco",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -254,7 +254,7 @@ blip_itm_base_coco = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/Salesforce/blip-itm-base-coco",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -266,7 +266,7 @@ blip_itm_base_coco = ModelMeta(
 )
 
 blip_itm_large_coco = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-itm-large-coco",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -281,7 +281,7 @@ blip_itm_large_coco = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/Salesforce/blip-itm-large-coco",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -294,7 +294,7 @@ blip_itm_large_coco = ModelMeta(
 )
 
 blip_itm_base_flickr = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-itm-base-flickr",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -309,7 +309,7 @@ blip_itm_base_flickr = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/Salesforce/blip-itm-base-flickr",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -322,7 +322,7 @@ blip_itm_base_flickr = ModelMeta(
 )
 
 blip_itm_large_flickr = ModelMeta(
-    loader=BLIPModel,  # type: ignore
+    loader=BLIPModel,
     name="Salesforce/blip-itm-large-flickr",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -337,7 +337,7 @@ blip_itm_large_flickr = ModelMeta(
     open_weights=True,
     public_training_code="https://github.com/salesforce/BLIP",
     public_training_data="https://github.com/salesforce/BLIP",
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/Salesforce/blip-itm-large-flickr",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,

mteb/models/model_implementations/bm25.py

@@ -1,5 +1,4 @@
 import logging
-from typing import Any
 
 from mteb._create_dataloaders import _create_text_queries_dataloader
 from mteb._requires_package import requires_package
@@ -8,6 +7,7 @@ from mteb.models.model_meta import ModelMeta
 from mteb.models.models_protocols import SearchProtocol
 from mteb.types import (
     CorpusDatasetType,
+    EncodeKwargs,
     InstructionDatasetType,
     QueryDatasetType,
     RetrievalOutputType,
@@ -49,7 +49,7 @@ def bm25_loader(model_name, **kwargs) -> SearchProtocol:
             task_metadata: TaskMetadata,
             hf_split: str,
             hf_subset: str,
-            encode_kwargs: dict[str, Any],
+            encode_kwargs: EncodeKwargs,
         ) -> None:
             logger.info("Encoding Corpus...")
             corpus_texts = [
@@ -74,7 +74,7 @@ def bm25_loader(model_name, **kwargs) -> SearchProtocol:
             hf_split: str,
             hf_subset: str,
             top_k: int,
-            encode_kwargs: dict[str, Any],
+            encode_kwargs: EncodeKwargs,
             instructions: InstructionDatasetType | None = None,
             top_ranked: TopRankedDocumentsType | None = None,
         ) -> RetrievalOutputType:
@@ -113,7 +113,7 @@ def bm25_loader(model_name, **kwargs) -> SearchProtocol:
 
         def encode(self, texts: list[str]):
             """Encode input text as term vectors"""
-            return bm25s.tokenize(texts, stopwords=self.stopwords, stemmer=self.stemmer)  # type: ignore
+            return bm25s.tokenize(texts, stopwords=self.stopwords, stemmer=self.stemmer)
 
     return BM25Search(**kwargs)
 

mteb/models/model_implementations/bmretriever_models.py

@@ -25,6 +25,7 @@ class BMRetrieverWrapper(InstructSentenceTransformerModel):
         self,
         model_name: str,
         revision: str,
+        device: str | None = None,
         instruction_template: str
         | Callable[[str, PromptType | None], str]
         | None = None,
@@ -52,6 +53,7 @@ class BMRetrieverWrapper(InstructSentenceTransformerModel):
 
         transformer = Transformer(
             model_name,
+            device=device,
             **kwargs,
         )
         pooling = Pooling(
@@ -102,7 +104,7 @@ BMRetriever_410M = ModelMeta(
     license="mit",
     reference="https://huggingface.co/BMRetriever/BMRetriever-410M",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -132,7 +134,7 @@ BMRetriever_1B = ModelMeta(
     license="mit",
     reference="https://huggingface.co/BMRetriever/BMRetriever-1B",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -162,7 +164,7 @@ BMRetriever_2B = ModelMeta(
     license="mit",
     reference="https://huggingface.co/BMRetriever/BMRetriever-2B",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,
@@ -192,7 +194,7 @@ BMRetriever_7B = ModelMeta(
     license="mit",
     reference="https://huggingface.co/BMRetriever/BMRetriever-7B",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "Transformers", "safetensors"],
     use_instructions=True,
     public_training_code=None,
     public_training_data=None,

mteb/models/model_implementations/cadet_models.py

@@ -47,7 +47,7 @@ cadet_embed = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/manveertamber/cadet-embed-base-v1",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors"],
     use_instructions=True,
     public_training_code="https://github.com/manveertamber/cadet-dense-retrieval",
     # we provide the code to generate the training data

mteb/models/model_implementations/cde_models.py

@@ -49,10 +49,17 @@ class CDEWrapper(SentenceTransformerEncoderWrapper):
         "InstructionReranking",
     )
 
-    def __init__(self, model: str, *args, **kwargs: Any) -> None:
+    def __init__(
+        self,
+        model: str,
+        revision: str | None = None,
+        device: str | None = None,
+        *args,
+        **kwargs: Any,
+    ) -> None:
         from transformers import AutoConfig
 
-        super().__init__(model, *args, **kwargs)
+        super().__init__(model, revision=revision, device=device, *args, **kwargs)
         model_config = AutoConfig.from_pretrained(model, trust_remote_code=True)
         self.max_sentences = model_config.transductive_corpus_size
 
@@ -220,7 +227,7 @@ cde_small_v1 = ModelMeta(
     embed_dim=768,
     license="mit",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers"],
+    framework=["Sentence Transformers", "safetensors", "Transformers"],
     reference="https://huggingface.co/jxm/cde-small-v1",
     use_instructions=True,
     adapted_from="nomic-ai/nomic-bert-2048",
@@ -249,7 +256,7 @@ cde_small_v2 = ModelMeta(
     embed_dim=768,
     license="mit",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers"],
+    framework=["Sentence Transformers", "safetensors", "Transformers"],
     reference="https://huggingface.co/jxm/cde-small-v1",
     use_instructions=True,
     adapted_from="answerdotai/ModernBERT-base",

mteb/models/model_implementations/clip_models.py

@@ -115,7 +115,7 @@ CLIP_CITATION = """
 
 
 clip_vit_large_patch14 = ModelMeta(
-    loader=CLIPModel,  # type: ignore
+    loader=CLIPModel,
     name="openai/clip-vit-large-patch14",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -130,7 +130,7 @@ clip_vit_large_patch14 = ModelMeta(
     open_weights=True,
     public_training_code=None,
     public_training_data=None,
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers", "safetensors"],
     reference="https://huggingface.co/openai/clip-vit-large-patch14",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -139,7 +139,7 @@ clip_vit_large_patch14 = ModelMeta(
 )
 
 clip_vit_base_patch32 = ModelMeta(
-    loader=CLIPModel,  # type: ignore
+    loader=CLIPModel,
     name="openai/clip-vit-base-patch32",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -154,7 +154,7 @@ clip_vit_base_patch32 = ModelMeta(
     open_weights=True,
     public_training_code=None,
     public_training_data=None,
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/openai/clip-vit-base-patch32",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
@@ -163,7 +163,7 @@ clip_vit_base_patch32 = ModelMeta(
 )
 
 clip_vit_base_patch16 = ModelMeta(
-    loader=CLIPModel,  # type: ignore
+    loader=CLIPModel,
     name="openai/clip-vit-base-patch16",
     model_type=["dense"],
     languages=["eng-Latn"],
@@ -178,7 +178,7 @@ clip_vit_base_patch16 = ModelMeta(
     open_weights=True,
     public_training_code=None,
     public_training_data=None,
-    framework=["PyTorch"],
+    framework=["PyTorch", "Transformers"],
     reference="https://huggingface.co/openai/clip-vit-base-patch16",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,

mteb/models/model_implementations/clips_models.py

@@ -36,7 +36,7 @@ e5_nl_small = ModelMeta(
     max_tokens=512,
     reference="https://huggingface.co/clips/e5-small-trm-nl",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=True,
     public_training_code="https://github.com/ELotfi/e5-nl",
     public_training_data="https://huggingface.co/collections/clips/beir-nl",
@@ -63,7 +63,7 @@ e5_nl_base = ModelMeta(
     max_tokens=514,
     reference="https://huggingface.co/clips/e5-base-trm-nl",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=True,
     public_training_code="https://github.com/ELotfi/e5-nl",
     public_training_data="https://huggingface.co/collections/clips/beir-nl",
@@ -90,7 +90,7 @@ e5_nl_large = ModelMeta(
     max_tokens=514,
     reference="https://huggingface.co/clips/e5-large-trm-nl",
     similarity_fn_name=ScoringFunction.COSINE,
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=True,
     public_training_code="https://github.com/ELotfi/e5-nl",
     public_training_data="https://huggingface.co/collections/clips/beir-nl",

mteb/models/model_implementations/codefuse_models.py

@@ -242,7 +242,7 @@ F2LLM_0B6 = ModelMeta(
     max_tokens=8192,
     reference="https://huggingface.co/codefuse-ai/F2LLM-0.6B",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=True,
     public_training_code="https://github.com/codefuse-ai/F2LLM",
     public_training_data="https://huggingface.co/datasets/codefuse-ai/F2LLM",
@@ -272,7 +272,7 @@ F2LLM_1B7 = ModelMeta(
     max_tokens=8192,
     reference="https://huggingface.co/codefuse-ai/F2LLM-1.7B",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=True,
     public_training_code="https://github.com/codefuse-ai/F2LLM",
     public_training_data="https://huggingface.co/datasets/codefuse-ai/F2LLM",
@@ -302,7 +302,7 @@ F2LLM_4B = ModelMeta(
     max_tokens=8192,
     reference="https://huggingface.co/codefuse-ai/F2LLM-4B",
     similarity_fn_name="cosine",
-    framework=["Sentence Transformers", "PyTorch"],
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
     use_instructions=True,
     public_training_code="https://github.com/codefuse-ai/F2LLM",
     public_training_data="https://huggingface.co/datasets/codefuse-ai/F2LLM",
@@ -325,7 +325,7 @@ C2LLM_0B5 = ModelMeta(
     open_weights=True,
     public_training_code=None,
     public_training_data=None,
-    framework=["PyTorch", "Sentence Transformers"],
+    framework=["PyTorch", "Sentence Transformers", "Transformers", "safetensors"],
     reference="https://huggingface.co/codefuse-ai/C2LLM-0.5B",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=True,
@@ -353,7 +353,7 @@ C2LLM_7B = ModelMeta(
     open_weights=True,
     public_training_code=None,
     public_training_data=None,
-    framework=["PyTorch", "Sentence Transformers"],
+    framework=["PyTorch", "Sentence Transformers", "Transformers", "safetensors"],
     reference="https://huggingface.co/codefuse-ai/C2LLM-7B",
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=True,