mteb 2.1.7__py3-none-any.whl → 2.1.8__py3-none-any.whl

mteb/models/model_implementations/blip2_models.py

@@ -10,6 +10,13 @@ from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta, ScoringFunction
 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},
+    author={Junnan Li and Dongxu Li and Silvio Savarese and Steven Hoi},
+    year={2023},
+    booktitle={ICML},
+}"""
+
 
 def blip2_loader(model_name, **kwargs):
     requires_package(
@@ -176,6 +183,7 @@ blip2_opt_2_7b = ModelMeta(
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
     training_datasets=blip2_training_datasets,
+    citation=BLIP2_CITATION,
 )
 
 blip2_opt_6_7b_coco = ModelMeta(
@@ -198,4 +206,5 @@ blip2_opt_6_7b_coco = ModelMeta(
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
     training_datasets=blip2_training_datasets,
+    citation=BLIP2_CITATION,
 )

mteb/models/model_implementations/blip_models.py

@@ -10,6 +10,17 @@ from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta, ScoringFunction
 from mteb.types import Array, BatchedInput, PromptType
 
+BLIP_CITATION = """@misc{https://doi.org/10.48550/arxiv.2201.12086,
+    doi = {10.48550/ARXIV.2201.12086},
+    url = {https://arxiv.org/abs/2201.12086},
+    author = {Li, Junnan and Li, Dongxu and Xiong, Caiming and Hoi, Steven},
+    keywords = {Computer Vision and Pattern Recognition (cs.CV), FOS: Computer and information sciences, FOS: Computer and information sciences},
+    title = {BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation},
+    publisher = {arXiv},
+    year = {2022},
+    copyright = {Creative Commons Attribution 4.0 International}
+}"""
+
 
 class BLIPModel(AbsEncoder):
     def __init__(
@@ -140,6 +151,7 @@ blip_image_captioning_large = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )
 
 blip_image_captioning_base = ModelMeta(
@@ -166,6 +178,7 @@ blip_image_captioning_base = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )
 
 
@@ -192,6 +205,7 @@ blip_vqa_base = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )
 
 blip_vqa_capfilt_large = ModelMeta(
@@ -217,6 +231,7 @@ blip_vqa_capfilt_large = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )
 
 blip_itm_base_coco = ModelMeta(
@@ -242,6 +257,7 @@ blip_itm_base_coco = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )
 
 blip_itm_large_coco = ModelMeta(
@@ -268,6 +284,7 @@ blip_itm_large_coco = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )
 
 blip_itm_base_flickr = ModelMeta(
@@ -294,6 +311,7 @@ blip_itm_base_flickr = ModelMeta(
         # LAION115M
         # Flickr30k
     ),
+    citation=BLIP_CITATION,
 )
 
 blip_itm_large_flickr = ModelMeta(
@@ -319,4 +337,5 @@ blip_itm_large_flickr = ModelMeta(
         # CC3M+CC12M+SBU
         # LAION115M
     ),
+    citation=BLIP_CITATION,
 )

mteb/models/model_implementations/cadet_models.py

@@ -3,6 +3,13 @@ from mteb.models.sentence_transformer_wrapper import sentence_transformers_loade
 
 from .bge_models import bge_m3_training_data
 
+CADET_CITATION = """@article{tamber2025conventionalcontrastivelearningfalls,
+    title={Conventional Contrastive Learning Often Falls Short: Improving Dense Retrieval with Cross-Encoder Listwise Distillation and Synthetic Data},
+    author={Manveer Singh Tamber and Suleman Kazi and Vivek Sourabh and Jimmy Lin},
+    journal={arXiv:2505.19274},
+    year={2025}
+}"""
+
 cadet_training_data = {
     # we train with the corpora of FEVER, MSMARCO, and DBPEDIA. We only train with synthetic generated queries.
     # However, we do use queries from MSMARCO as examples for synthetic query generation.
@@ -46,4 +53,5 @@ cadet_embed = ModelMeta(
     public_training_data="https://github.com/manveertamber/cadet-dense-retrieval",
     training_datasets=cadet_training_data,
     adapted_from="intfloat/e5-base-unsupervised",
+    citation=CADET_CITATION,
 )

mteb/models/model_implementations/cde_models.py

@@ -24,6 +24,16 @@ if TYPE_CHECKING:
     )
 logger = logging.getLogger(__name__)
 
+CDE_CITATION = """@misc{morris2024contextualdocumentembeddings,
+    title={Contextual Document Embeddings},
+    author={John X. Morris and Alexander M. Rush},
+    year={2024},
+    eprint={2410.02525},
+    archivePrefix={arXiv},
+    primaryClass={cs.CL},
+    url={https://arxiv.org/abs/2410.02525},
+}"""
+
 
 class CDEWrapper(SentenceTransformerEncoderWrapper):
     dataset_embeddings: torch.Tensor | None = None
@@ -217,6 +227,7 @@ cde_small_v1 = ModelMeta(
     training_datasets=bge_full_data,
     public_training_code="https://github.com/jxmorris12/cde",
     public_training_data="https://huggingface.co/datasets/cfli/bge-full-data",
+    citation=CDE_CITATION,
 )
 
 cde_small_v2 = ModelMeta(
@@ -244,4 +255,5 @@ cde_small_v2 = ModelMeta(
     training_datasets=bge_full_data,
     public_training_code="https://github.com/jxmorris12/cde",
     public_training_data="https://huggingface.co/datasets/cfli/bge-full-data",
+    citation=CDE_CITATION,
 )

mteb/models/model_implementations/codefuse_models.py

@@ -2,6 +2,18 @@ from mteb.models import ModelMeta
 from mteb.models.instruct_wrapper import InstructSentenceTransformerModel
 from mteb.types import PromptType
 
+F2LLM_CITATION = """@article{2025F2LLM,
+    title={F2LLM Technical Report: Matching SOTA Embedding Performance with 6 Million Open-Source Data},
+    author={Ziyin Zhang and Zihan Liao and Hang Yu and Peng Di and Rui Wang},
+    journal={CoRR},
+    volume={abs/2510.02294},
+    year={2025},
+    url={https://doi.org/10.48550/arXiv.2510.02294},
+    doi={10.48550/ARXIV.2510.02294},
+    eprinttype={arXiv},
+    eprint={2510.02294}
+}"""
+
 training_datasets = {
     "MSMARCO",
     "ArguAna",
@@ -146,6 +158,7 @@ F2LLM_0B6 = ModelMeta(
     public_training_code="https://github.com/codefuse-ai/F2LLM",
     public_training_data="https://huggingface.co/datasets/codefuse-ai/F2LLM",
     training_datasets=training_datasets,
+    citation=F2LLM_CITATION,
 )
 
 F2LLM_1B7 = ModelMeta(
@@ -174,6 +187,7 @@ F2LLM_1B7 = ModelMeta(
     public_training_code="https://github.com/codefuse-ai/F2LLM",
     public_training_data="https://huggingface.co/datasets/codefuse-ai/F2LLM",
     training_datasets=training_datasets,
+    citation=F2LLM_CITATION,
 )
 
 F2LLM_4B = ModelMeta(
@@ -202,4 +216,5 @@ F2LLM_4B = ModelMeta(
     public_training_code="https://github.com/codefuse-ai/F2LLM",
     public_training_data="https://huggingface.co/datasets/codefuse-ai/F2LLM",
     training_datasets=training_datasets,
+    citation=F2LLM_CITATION,
 )

mteb/models/model_implementations/codesage_models.py

@@ -1,6 +1,15 @@
 from mteb.models.model_meta import ModelMeta, ScoringFunction
 from mteb.models.sentence_transformer_wrapper import sentence_transformers_loader
 
+CODESAGE_CITATION = """@inproceedings{
+    zhang2024code,
+    title={{CODE} {REPRESENTATION} {LEARNING} {AT} {SCALE}},
+    author={Dejiao Zhang and Wasi Uddin Ahmad and Ming Tan and Hantian Ding and Ramesh Nallapati and Dan Roth and Xiaofei Ma and Bing Xiang},
+    booktitle={The Twelfth International Conference on Learning Representations},
+    year={2024},
+    url={https://openreview.net/forum?id=vfzRRjumpX}
+}"""
+
 codesage_languages = [
     "python-Code",
     "javascript-Code",
@@ -33,6 +42,7 @@ codesage_large = ModelMeta(
         "CodeSearchNetRetrieval",
         "CodeSearchNetCCRetrieval",
     },
+    citation=CODESAGE_CITATION,
 )
 
 codesage_base = ModelMeta(
@@ -58,6 +68,7 @@ codesage_base = ModelMeta(
         "CodeSearchNetRetrieval",
         "CodeSearchNetCCRetrieval",
     },
+    citation=CODESAGE_CITATION,
 )
 
 codesage_small = ModelMeta(
@@ -83,4 +94,5 @@ codesage_small = ModelMeta(
         "CodeSearchNetRetrieval",
         "CodeSearchNetCCRetrieval",
     },
+    citation=CODESAGE_CITATION,
 )

mteb/models/model_implementations/misc_models.py

@@ -31,6 +31,12 @@ Haon_Chen__speed_embedding_7b_instruct = ModelMeta(
     training_datasets=None,
     adapted_from="mistralai/Mistral-7B-v0.1",
     superseded_by=None,
+    citation="""@article{chen2024little,
+    title={Little Giants: Synthesizing High-Quality Embedding Data at Scale},
+    author={Chen, Haonan and Wang, Liang and Yang, Nan and Zhu, Yutao and Zhao, Ziliang and Wei, Furu and Dou, Zhicheng},
+    journal={arXiv preprint arXiv:2410.18634},
+    year={2024}
+}""",
 )
 Gameselo__STS_multilingual_mpnet_base_v2 = ModelMeta(
     name="Gameselo/STS-multilingual-mpnet-base-v2",

mteb/models/model_implementations/moco_models.py

@@ -10,6 +10,13 @@ from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta, ScoringFunction
 from mteb.types import Array, BatchedInput, PromptType
 
+MOCOV3_CITATION = """@Article{chen2021mocov3,
+    author  = {Xinlei Chen* and Saining Xie* and Kaiming He},
+    title   = {An Empirical Study of Training Self-Supervised Vision Transformers},
+    journal = {arXiv preprint arXiv:2104.02057},
+    year    = {2021},
+}"""
+
 
 def mocov3_loader(model_name, **kwargs):
     requires_package(mocov3_loader, "timm", model_name, "pip install 'mteb[timm]'")
@@ -129,6 +136,7 @@ mocov3_vit_base = ModelMeta(
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
     training_datasets=mocov3_training_datasets,
+    citation=MOCOV3_CITATION,
 )
 
 mocov3_vit_large = ModelMeta(
@@ -151,4 +159,5 @@ mocov3_vit_large = ModelMeta(
     similarity_fn_name=ScoringFunction.COSINE,
     use_instructions=False,
     training_datasets=mocov3_training_datasets,
+    citation=MOCOV3_CITATION,
 )

mteb/models/model_implementations/openclip_models.py

@@ -10,6 +10,14 @@ from mteb.models.abs_encoder import AbsEncoder
 from mteb.models.model_meta import ModelMeta, ScoringFunction
 from mteb.types import Array, BatchedInput, PromptType
 
+OPENCLIP_CITATION = """@inproceedings{cherti2023reproducible,
+    title={Reproducible scaling laws for contrastive language-image learning},
+    author={Cherti, Mehdi and Beaumont, Romain and Wightman, Ross and Wortsman, Mitchell and Ilharco, Gabriel and Gordon, Cade and Schuhmann, Christoph and Schmidt, Ludwig and Jitsev, Jenia},
+    booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
+    pages={2818--2829},
+    year={2023}
+}"""
+
 
 def openclip_loader(model_name, **kwargs):
     requires_package(
@@ -133,6 +141,7 @@ CLIP_ViT_L_14_DataComp_XL_s13B_b90K = ModelMeta(
     training_datasets=set(
         # DataComp-1B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_B_32_DataComp_XL_s13B_b90K = ModelMeta(
@@ -157,6 +166,7 @@ CLIP_ViT_B_32_DataComp_XL_s13B_b90K = ModelMeta(
     training_datasets=set(
         # DataComp-1B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_B_16_DataComp_XL_s13B_b90K = ModelMeta(
@@ -181,6 +191,7 @@ CLIP_ViT_B_16_DataComp_XL_s13B_b90K = ModelMeta(
     training_datasets=set(
         # DataComp-1B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_bigG_14_laion2B_39B_b160k = ModelMeta(
@@ -205,6 +216,7 @@ CLIP_ViT_bigG_14_laion2B_39B_b160k = ModelMeta(
     training_datasets=set(
         # 2 Billion sample English subset of LAION-5B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_g_14_laion2B_s34B_b88K = ModelMeta(
@@ -229,6 +241,7 @@ CLIP_ViT_g_14_laion2B_s34B_b88K = ModelMeta(
     training_datasets=set(
         # 2 Billion sample English subset of LAION-5B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_H_14_laion2B_s32B_b79K = ModelMeta(
@@ -253,6 +266,7 @@ CLIP_ViT_H_14_laion2B_s32B_b79K = ModelMeta(
     training_datasets=set(
         # 2 Billion sample English subset of LAION-5B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_L_14_laion2B_s32B_b82K = ModelMeta(
@@ -277,6 +291,7 @@ CLIP_ViT_L_14_laion2B_s32B_b82K = ModelMeta(
     training_datasets=set(
         # 2 Billion sample English subset of LAION-5B
     ),
+    citation=OPENCLIP_CITATION,
 )
 
 CLIP_ViT_B_32_laion2B_s34B_b79K = ModelMeta(
@@ -301,4 +316,5 @@ CLIP_ViT_B_32_laion2B_s34B_b79K = ModelMeta(
     training_datasets=set(
         # 2 Billion sample English subset of LAION-5B
     ),
+    citation=OPENCLIP_CITATION,
 )

mteb/models/model_implementations/piccolo_models.py

@@ -48,4 +48,10 @@ piccolo_large_zh_v2 = ModelMeta(
     public_training_code=None,
     public_training_data=None,
     training_datasets=None,  # They don't say
+    citation="""@misc{2405.06932,
+    Author = {Junqin Huang and Zhongjie Hu and Zihao Jing and Mengya Gao and Yichao Wu},
+    Title = {Piccolo2: General Text Embedding with Multi-task Hybrid Loss Training},
+    Year = {2024},
+    Eprint = {arXiv:2405.06932},
+}""",
 )

mteb/models/model_implementations/rasgaard_models.py

@@ -17,11 +17,17 @@ potion_base_8m = ModelMeta(
     license="mit",
     similarity_fn_name=ScoringFunction.COSINE,
     framework=["NumPy", "Sentence Transformers"],
-    reference="https://huggingface.co/minishlab/rasgaard/m2v-dfm-large",
+    reference="https://huggingface.co/rasgaard/m2v-dfm-large",
     use_instructions=False,
     adapted_from="KennethEnevoldsen/dfm-sentence-encoder-large",
     superseded_by=None,
     training_datasets=set(),  # distilled
     public_training_code="https://github.com/MinishLab/model2vec",
     public_training_data="https://huggingface.co/datasets/HuggingFaceFW/fineweb-2",  # distilled on this
+    citation="""@article{minishlab2024model2vec,
+    author = {Tulkens, Stephan and {van Dongen}, Thomas},
+    title = {Model2Vec: Fast State-of-the-Art Static Embeddings},
+    year = {2024},
+    url = {https://github.com/MinishLab/model2vec}
+}""",
 )