mteb 2.3.2__tar.gz → 2.3.3__tar.gz

{mteb-2.3.2 → mteb-2.3.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mteb
-Version: 2.3.2
+Version: 2.3.3
 Summary: Massive Text Embedding Benchmark
 Author-email: MTEB Contributors <niklas@huggingface.co>, Kenneth Enevoldsen <kenneth.enevoldsen@cas.au.dk>, Nouamane Tazi <nouamane@huggingface.co>, Nils Reimers <info@nils-reimers.de>
 Maintainer-email: Kenneth Enevoldsen <kenneth.enevoldsen@cas.au.dk>, Roman Solomatin <risolomatin@gmail.com>, Isaac Chung <chungisaac1217@gmail.com>
@@ -108,7 +108,7 @@ Requires-Dist: qwen_vl_utils>=0.0.14; extra == "eager-embed"
 Dynamic: license-file
 
 <h1 align="center">
-  <img src="docs/images/logos/mteb_logo/dots-icon.png" alt="MTEB" width="28" style="vertical-align: middle; margin-right: 10px;"/> MTEB
+  <img src="https://github.com/embeddings-benchmark/mteb/blob/main/docs/images/logos/mteb_logo/dots-icon.png?raw=true" alt="MTEB" width="28" style="vertical-align: middle; margin-right: 10px;"/> MTEB
 </h1>
 
 <h3 align="center" style="border-bottom: none;">Multimodal toolbox for evaluating embeddings and retrieval systems</h3>
@@ -137,7 +137,7 @@ Dynamic: license-file
 
 
 <h3 align="center">
-    <a href="https://huggingface.co/spaces/mteb/leaderboard"><img style="float: middle; padding: 10px 10px 10px 10px;" width="60" height="55" src="./docs/images/logos/hf_logo.png" /></a>
+    <a href="https://huggingface.co/spaces/mteb/leaderboard"><img style="float: middle; padding: 10px 10px 10px 10px;" width="60" height="55" src="https://github.com/embeddings-benchmark/mteb/blob/main/docs/images/logos/hf_logo.png?raw=true" /></a>
 </h3>
 
 

{mteb-2.3.2 → mteb-2.3.3}/README.md

@@ -1,5 +1,5 @@
 <h1 align="center">
-  <img src="docs/images/logos/mteb_logo/dots-icon.png" alt="MTEB" width="28" style="vertical-align: middle; margin-right: 10px;"/> MTEB
+  <img src="https://github.com/embeddings-benchmark/mteb/blob/main/docs/images/logos/mteb_logo/dots-icon.png?raw=true" alt="MTEB" width="28" style="vertical-align: middle; margin-right: 10px;"/> MTEB
 </h1>
 
 <h3 align="center" style="border-bottom: none;">Multimodal toolbox for evaluating embeddings and retrieval systems</h3>
@@ -28,7 +28,7 @@
 
 
 <h3 align="center">
-    <a href="https://huggingface.co/spaces/mteb/leaderboard"><img style="float: middle; padding: 10px 10px 10px 10px;" width="60" height="55" src="./docs/images/logos/hf_logo.png" /></a>
+    <a href="https://huggingface.co/spaces/mteb/leaderboard"><img style="float: middle; padding: 10px 10px 10px 10px;" width="60" height="55" src="https://github.com/embeddings-benchmark/mteb/blob/main/docs/images/logos/hf_logo.png?raw=true" /></a>
 </h3>
 
 

{mteb-2.3.2 → mteb-2.3.3}/mteb/benchmarks/_create_table.py

@@ -1,4 +1,3 @@
-import math
 import re
 from collections import defaultdict
 
@@ -32,26 +31,18 @@ def _split_on_capital(s: str) -> str:
     return " ".join(re.findall(r"[A-Z]?[a-z]+|[A-Z]+(?=[A-Z]|$)", s))
 
 
-def _format_n_parameters(n_parameters) -> str:
-    if (n_parameters is None) or (not int(n_parameters)):
-        return "Unknown"
-    n_thousand = int(n_parameters // 1e3)
-    if n_thousand < 1:
-        return str(int(n_parameters))
-    n_zeros = math.log10(n_thousand)
-    if n_zeros >= 6:
-        return str(n_thousand // (10**6)) + "B"
-    if n_zeros >= 3:
-        return str(n_thousand // (10**3)) + "M"
-    return str(n_thousand) + "K"
+def _format_n_parameters(n_parameters) -> float | None:
+    """Format n_parameters to be in billions with decimals down to 1 million. I.e. 7M -> 0.007B, 1.5B -> 1.5B, None -> None"""
+    if n_parameters:
+        n_parameters = float(n_parameters)
+        return round(n_parameters / 1e9, 3)
+    return None
 
 
-def _format_max_tokens(max_tokens: float | None) -> str:
-    if max_tokens is None:
-        return "Unknown"
-    if max_tokens == np.inf:
-        return "Infinite"
-    return str(int(max_tokens))
+def _format_max_tokens(max_tokens: float | None) -> float | None:
+    if max_tokens is None or max_tokens == np.inf:
+        return None
+    return float(max_tokens)
 
 
 def _get_means_per_types(per_task: pd.DataFrame):
@@ -144,18 +135,18 @@ def _create_summary_table_from_benchmark_results(
     joint_table.insert(
         1,
         "Embedding Dimensions",
-        model_metas.map(lambda m: str(int(m.embed_dim)) if m.embed_dim else "Unknown"),
+        model_metas.map(lambda m: int(m.embed_dim) if m.embed_dim else None),
     )
     joint_table.insert(
         1,
-        "Number of Parameters",
+        "Number of Parameters (B)",
         model_metas.map(lambda m: _format_n_parameters(m.n_parameters)),
     )
     joint_table.insert(
         1,
         "Memory Usage (MB)",
         model_metas.map(
-            lambda m: str(int(m.memory_usage_mb)) if m.memory_usage_mb else "Unknown"
+            lambda m: int(m.memory_usage_mb) if m.memory_usage_mb else None
         ),
     )
 
@@ -323,18 +314,18 @@ def _create_summary_table_mean_public_private(
     joint_table.insert(
         1,
         "Embedding Dimensions",
-        model_metas.map(lambda m: str(int(m.embed_dim)) if m.embed_dim else "Unknown"),
+        model_metas.map(lambda m: int(m.embed_dim) if m.embed_dim else None),
     )
     joint_table.insert(
         1,
-        "Number of Parameters",
+        "Number of Parameters (B)",
         model_metas.map(lambda m: _format_n_parameters(m.n_parameters)),
     )
     joint_table.insert(
         1,
         "Memory Usage (MB)",
         model_metas.map(
-            lambda m: str(int(m.memory_usage_mb)) if m.memory_usage_mb else "Unknown"
+            lambda m: int(m.memory_usage_mb) if m.memory_usage_mb else None
         ),
     )
 
@@ -445,18 +436,18 @@ def _create_summary_table_mean_subset(
     joint_table.insert(
         1,
         "Embedding Dimensions",
-        model_metas.map(lambda m: str(int(m.embed_dim)) if m.embed_dim else "Unknown"),
+        model_metas.map(lambda m: int(m.embed_dim) if m.embed_dim else None),
     )
     joint_table.insert(
         1,
-        "Number of Parameters",
+        "Number of Parameters (B)",
         model_metas.map(lambda m: _format_n_parameters(m.n_parameters)),
     )
     joint_table.insert(
         1,
         "Memory Usage (MB)",
         model_metas.map(
-            lambda m: str(int(m.memory_usage_mb)) if m.memory_usage_mb else "Unknown"
+            lambda m: int(m.memory_usage_mb) if m.memory_usage_mb else None
         ),
     )
 
@@ -558,25 +549,23 @@ def _create_summary_table_mean_task_type(
 
     # Insert model metadata columns
     joint_table.insert(
-        1,
-        "Max Tokens",
-        model_metas.map(lambda m: _format_max_tokens(m.max_tokens)),
+        1, "Max Tokens", model_metas.map(lambda m: _format_max_tokens(m.max_tokens))
     )
     joint_table.insert(
         1,
         "Embedding Dimensions",
-        model_metas.map(lambda m: str(int(m.embed_dim)) if m.embed_dim else "Unknown"),
+        model_metas.map(lambda m: int(m.embed_dim) if m.embed_dim else None),
     )
     joint_table.insert(
         1,
-        "Number of Parameters",
+        "Number of Parameters (B)",
         model_metas.map(lambda m: _format_n_parameters(m.n_parameters)),
     )
     joint_table.insert(
         1,
         "Memory Usage (MB)",
         model_metas.map(
-            lambda m: str(int(m.memory_usage_mb)) if m.memory_usage_mb else "Unknown"
+            lambda m: int(m.memory_usage_mb) if m.memory_usage_mb else None
         ),
     )
 

{mteb-2.3.2 → mteb-2.3.3}/mteb/leaderboard/app.py

@@ -535,9 +535,6 @@ def get_leaderboard_app(cache: ResultCache = ResultCache()) -> gr.Blocks:
 
         with gr.Tab("Performance per Model Size") as plot_tab:
             plot = gr.Plot(_performance_size_plot, inputs=[summary_table])
-            gr.Markdown(
-                "*We only display TOP 5 models that have been run on all tasks in the benchmark*"
-            )
             plot_tab.select(
                 _performance_size_plot, inputs=[summary_table], outputs=[plot]
             )

{mteb-2.3.2 → mteb-2.3.3}/mteb/leaderboard/figures.py

@@ -1,3 +1,4 @@
+import logging
 from typing import get_args
 
 import numpy as np
@@ -7,6 +8,8 @@ import plotly.graph_objects as go
 
 from mteb.abstasks.task_metadata import TaskType
 
+logger = logging.getLogger(__name__)
+
 
 def _text_plot(text: str):
     """Returns empty scatter plot with text added, this can be great for error messages."""
@@ -29,16 +32,17 @@ def _failsafe_plot(fun):
         try:
             return fun(*args, **kwargs)
         except Exception as e:
+            logger.error(f"Plot generation failed: {e}")
             return _text_plot(f"Couldn't produce plot. Reason: {e}")
 
     return wrapper
 
 
-def _parse_n_params(text: str) -> int:
-    if text.endswith("M"):
-        return float(text[:-1]) * 1e6
-    if text.endswith("B"):
-        return float(text[:-1]) * 1e9
+def _parse_n_params(params: float | None) -> int | float:
+    """Specified in billions."""
+    if params is None or np.isnan(params):
+        return None
+    return int(params * 1e9)
 
 
 def _parse_model_name(name: str) -> str:
@@ -51,20 +55,14 @@ def _parse_model_name(name: str) -> str:
 
 
 def _parse_float(value) -> float:
-    try:
-        if value == "Infinite":
-            return np.inf
-        else:
-            return float(value)
-    except ValueError:
+    if value is None or np.isnan(value):
         return np.nan
+    return float(value)
 
 
 def _process_max_tokens(x):
-    if pd.isna(x):
+    if pd.isna(x) or x is None or np.isinf(x):
         return "Unknown"
-    if np.isinf(x):
-        return "Infinite"
     return str(int(x))
 
 
@@ -112,7 +110,7 @@ def _add_size_guide(fig: go.Figure):
 @_failsafe_plot
 def _performance_size_plot(df: pd.DataFrame) -> go.Figure:
     df = df.copy()
-    df["Number of Parameters"] = df["Number of Parameters"].map(_parse_n_params)
+    df["Number of Parameters"] = df["Number of Parameters (B)"].map(_parse_n_params)
     df["Model"] = df["Model"].map(_parse_model_name)
     df["model_text"] = df["Model"].where(df["Model"].isin(models_to_annotate), "")
     df["Embedding Dimensions"] = df["Embedding Dimensions"].map(_parse_float)

{mteb-2.3.2 → mteb-2.3.3}/mteb/leaderboard/table.py

@@ -120,6 +120,14 @@ def apply_per_task_styling_from_benchmark(
     return _apply_per_task_table_styling(per_task_df)
 
 
+def _style_number_of_parameters(num_params: float) -> str:
+    """Anything bigger than 1B is shown in billions with 1 decimal (e.g. 1.712 > 1.7) while anything smaller as 0.xxx B (e.g. 0.345 remains 0.345)"""
+    if num_params >= 1:
+        return f"{num_params:.1f}"
+    else:
+        return f"{num_params:.3f}"
+
+
 def _apply_summary_table_styling(joint_table: pd.DataFrame) -> gr.DataFrame:
     """Apply styling to a raw summary DataFrame
 
@@ -130,7 +138,7 @@ def _apply_summary_table_styling(joint_table: pd.DataFrame) -> gr.DataFrame:
         "Rank (Borda)",
         "Rank",
         "Model",
-        "Number of Parameters",
+        "Number of Parameters (B)",
         "Embedding Dimensions",
         "Max Tokens",
         "Memory Usage (MB)",
@@ -156,7 +164,14 @@ def _apply_summary_table_styling(joint_table: pd.DataFrame) -> gr.DataFrame:
     joint_table[score_columns] = joint_table[score_columns].map(_format_scores)
 
     joint_table_style = joint_table.style.format(
-        {**dict.fromkeys(score_columns, "{:.2f}"), "Rank (Borda)": "{:.0f}"},
+        {
+            **dict.fromkeys(score_columns, "{:.2f}"),
+            "Rank (Borda)": "{:.0f}",
+            "Memory Usage (MB)": "{:.0f}",
+            "Embedding Dimensions": "{:.0f}",
+            "Max Tokens": "{:.0f}",
+            "Number of Parameters (B)": lambda x: _style_number_of_parameters(x),
+        },
         na_rep="",
     )
     joint_table_style = joint_table_style.highlight_min(

mteb-2.3.3/mteb/models/model_implementations/clips_models.py

@@ -0,0 +1,97 @@
+from mteb.models.model_meta import (
+    ModelMeta,
+    ScoringFunction,
+)
+from mteb.models.sentence_transformer_wrapper import sentence_transformers_loader
+
+from .e5_models import ME5_TRAINING_DATA, model_prompts
+
+E5_NL_CITATION = """
+@misc{banar2025mtebnle5nlembeddingbenchmark,
+  archiveprefix = {arXiv},
+  author = {Nikolay Banar and Ehsan Lotfi and Jens Van Nooten and Cristina Arhiliuc and Marija Kliocaite and Walter Daelemans},
+  eprint = {2509.12340},
+  primaryclass = {cs.CL},
+  title = {MTEB-NL and E5-NL: Embedding Benchmark and Models for Dutch},
+  url = {https://arxiv.org/abs/2509.12340},
+  year = {2025},
+}
+"""
+
+e5_nl_small = ModelMeta(
+    loader=sentence_transformers_loader,
+    loader_kwargs=dict(
+        model_prompts=model_prompts,
+    ),
+    name="clips/e5-small-trm-nl",
+    languages=["nld-Latn"],
+    open_weights=True,
+    revision="0243664a6c5e12eef854b091eb283e51833c3e9f",
+    release_date="2025-09-23",
+    n_parameters=40_800_000,
+    memory_usage_mb=78,
+    embed_dim=384,
+    license="mit",
+    max_tokens=512,
+    reference="https://huggingface.co/clips/e5-small-trm-nl",
+    similarity_fn_name=ScoringFunction.COSINE,
+    framework=["Sentence Transformers", "PyTorch"],
+    use_instructions=True,
+    public_training_code="https://github.com/ELotfi/e5-nl",
+    public_training_data="https://huggingface.co/collections/clips/beir-nl",
+    training_datasets=ME5_TRAINING_DATA,  # mMARCO-NL, HotpotQA-NL, FEVER-NL, and LLM generated data
+    adapted_from="intfloat/multilingual-e5-small",
+    citation=E5_NL_CITATION,
+)
+
+e5_nl_base = ModelMeta(
+    loader=sentence_transformers_loader,
+    loader_kwargs=dict(
+        model_prompts=model_prompts,
+    ),
+    name="clips/e5-base-trm-nl",
+    languages=["nld-Latn"],
+    open_weights=True,
+    revision="6bd5722f236da48b4b8bcb28cc1fc478f7089956",
+    release_date="2025-09-23",
+    n_parameters=124_400_000,
+    memory_usage_mb=237,
+    embed_dim=768,
+    license="mit",
+    max_tokens=514,
+    reference="https://huggingface.co/clips/e5-base-trm-nl",
+    similarity_fn_name=ScoringFunction.COSINE,
+    framework=["Sentence Transformers", "PyTorch"],
+    use_instructions=True,
+    public_training_code="https://github.com/ELotfi/e5-nl",
+    public_training_data="https://huggingface.co/collections/clips/beir-nl",
+    adapted_from="intfloat/multilingual-e5-base",
+    training_datasets=ME5_TRAINING_DATA,  # mMARCO-NL, HotpotQA-NL, FEVER-NL, and LLM generated data
+    citation=E5_NL_CITATION,
+)
+
+e5_nl_large = ModelMeta(
+    loader=sentence_transformers_loader,
+    loader_kwargs=dict(
+        model_prompts=model_prompts,
+    ),
+    name="clips/e5-large-trm-nl",
+    languages=["nld-Latn"],
+    open_weights=True,
+    revision="683333f86ed9eb3699b5567f0fdabeb958d412b0",
+    release_date="2025-09-23",
+    n_parameters=355_000_000,
+    memory_usage_mb=1355,
+    embed_dim=1024,
+    license="mit",
+    max_tokens=514,
+    reference="https://huggingface.co/clips/e5-large-trm-nl",
+    similarity_fn_name=ScoringFunction.COSINE,
+    framework=["Sentence Transformers", "PyTorch"],
+    use_instructions=True,
+    public_training_code="https://github.com/ELotfi/e5-nl",
+    public_training_data="https://huggingface.co/collections/clips/beir-nl",
+    training_datasets=ME5_TRAINING_DATA,  # mMARCO-NL, HotpotQA-NL, FEVER-NL, and LLM generated data
+    adapted_from="intfloat/multilingual-e5-large",
+    citation=E5_NL_CITATION,
+)

{mteb-2.3.2 → mteb-2.3.3}/mteb/models/model_implementations/cohere_models.py

@@ -8,6 +8,7 @@ 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
@@ -219,6 +220,8 @@ class CohereTextEmbeddingModel(AbsEncoder):
         output_dimension: int | None = None,
         **kwargs,
     ) -> None:
+        requires_package(self, "cohere", model_name, "pip install 'mteb[cohere]'")
+
         import cohere  # type: ignore
 
         self.model_name = model_name.removeprefix("Cohere/Cohere-")

{mteb-2.3.2 → mteb-2.3.3}/mteb/models/model_implementations/google_models.py

@@ -147,7 +147,6 @@ class GoogleTextEmbeddingModel(AbsEncoder):
 google_text_emb_004 = ModelMeta(
     loader=GoogleTextEmbeddingModel,  # type: ignore[call-arg]
     loader_kwargs=dict(
-        model_name="text-embedding-004",
         model_prompts=MODEL_PROMPTS,
     ),
     name="google/text-embedding-004",
@@ -172,7 +171,6 @@ google_text_emb_004 = ModelMeta(
 google_text_emb_005 = ModelMeta(
     loader=GoogleTextEmbeddingModel,  # type: ignore[call-arg]
     loader_kwargs=dict(
-        model_name="text-embedding-005",
         model_prompts=MODEL_PROMPTS,
     ),
     name="google/text-embedding-005",
@@ -197,7 +195,6 @@ google_text_emb_005 = ModelMeta(
 google_text_multilingual_emb_002 = ModelMeta(
     loader=GoogleTextEmbeddingModel,  # type: ignore[call-arg]
     loader_kwargs=dict(
-        model_name="text-embedding-002",
         model_prompts=MODEL_PROMPTS,
     ),
     name="google/text-multilingual-embedding-002",
@@ -222,7 +219,6 @@ google_text_multilingual_emb_002 = ModelMeta(
 google_gemini_embedding_001 = ModelMeta(
     loader=GoogleTextEmbeddingModel,  # type: ignore[call-arg]
     loader_kwargs=dict(
-        model_name="gemini-embedding-001",
         model_prompts=MODEL_PROMPTS,
     ),
     name="google/gemini-embedding-001",

mteb-2.3.3/mteb/models/model_implementations/kennethenevoldsen_models.py

@@ -0,0 +1,72 @@
+from mteb.models.model_meta import ModelMeta, ScoringFunction
+from mteb.models.sentence_transformer_wrapper import (
+    sentence_transformers_loader,
+)
+
+dfm_enc_large = ModelMeta(
+    loader=sentence_transformers_loader,  # type: ignore
+    name="KennethEnevoldsen/dfm-sentence-encoder-large",
+    languages=["dan-Latn"],
+    open_weights=True,
+    revision="132c53391e7a780dc6a2f9a03724d0158fe7122c",
+    release_date="2023-07-12",
+    n_parameters=355087360,
+    memory_usage_mb=1554,
+    embed_dim=1024,
+    license="mit",
+    max_tokens=512,
+    reference="https://huggingface.co/KennethEnevoldsen/dfm-sentence-encoder-large",
+    similarity_fn_name=ScoringFunction.COSINE,
+    framework=["Sentence Transformers", "PyTorch"],
+    use_instructions=False,
+    superseded_by=None,
+    adapted_from="chcaa/dfm-encoder-large-v1",
+    training_datasets=set(),  # just contrastive pre-training
+    public_training_code="https://huggingface.co/KennethEnevoldsen/dfm-sentence-encoder-large#hyperparameters",
+    citation="""@article{enevoldsenScandinavianEmbeddingBenchmarks2024,
+    title = {The {Scandinavian} {Embedding} {Benchmarks}: {Comprehensive} {Assessment} of {Multilingual} and {Monolingual} {Text} {Embedding}},
+    shorttitle = {The {Scandinavian} {Embedding} {Benchmarks}},
+    url = {https://openreview.net/forum?id=pJl_i7HIA72},
+    language = {en},
+    urldate = {2024-04-12},
+    author = {Enevoldsen, Kenneth and Kardos, Márton and Muennighoff, Niklas and Nielbo, Kristoffer},
+    month = feb,
+    year = {2024},
+}
+""",
+    public_training_data="https://huggingface.co/datasets/danish-foundation-models/danish-gigaword",  # paragraphs extracted from Danish Gigaword
+)
+
+dfm_enc_med = ModelMeta(
+    loader=sentence_transformers_loader,  # type: ignore
+    name="KennethEnevoldsen/dfm-sentence-encoder-medium",
+    languages=["dan-Latn"],
+    open_weights=True,
+    revision="701bce95d499fa97610d57e8823c54fd1fb79930",
+    release_date="2023-07-12",
+    n_parameters=124445952,
+    memory_usage_mb=475,
+    embed_dim=768,
+    license="mit",
+    max_tokens=512,
+    reference="https://huggingface.co/KennethEnevoldsen/dfm-sentence-encoder-medium",
+    similarity_fn_name=ScoringFunction.COSINE,
+    framework=["Sentence Transformers", "PyTorch"],
+    use_instructions=False,
+    superseded_by=None,
+    adapted_from=None,
+    public_training_code=None,
+    training_datasets=set(),  # just contrastive pre-training
+    citation="""@article{enevoldsenScandinavianEmbeddingBenchmarks2024,
+    title = {The {Scandinavian} {Embedding} {Benchmarks}: {Comprehensive} {Assessment} of {Multilingual} and {Monolingual} {Text} {Embedding}},
+    shorttitle = {The {Scandinavian} {Embedding} {Benchmarks}},
+    url = {https://openreview.net/forum?id=pJl_i7HIA72},
+    language = {en},
+    urldate = {2024-04-12},
+    author = {Enevoldsen, Kenneth and Kardos, Márton and Muennighoff, Niklas and Nielbo, Kristoffer},
+    month = feb,
+    year = {2024},
+}
+""",
+    public_training_data=None,
+)

{mteb-2.3.2 → mteb-2.3.3}/mteb/models/model_implementations/linq_models.py

@@ -1,7 +1,7 @@
 import torch
 
+from mteb.models.instruct_wrapper import instruct_wrapper
 from mteb.models.model_meta import ModelMeta, ScoringFunction
-from mteb.models.sentence_transformer_wrapper import SentenceTransformerEncoderWrapper
 from mteb.types import PromptType
 
 from .e5_instruct import E5_MISTRAL_TRAINING_DATA
@@ -22,7 +22,7 @@ def instruction_template(
 
 
 Linq_Embed_Mistral = ModelMeta(
-    loader=SentenceTransformerEncoderWrapper,
+    loader=instruct_wrapper,
     loader_kwargs=dict(
         instruction_template=instruction_template,
         attn="cccc",

{mteb-2.3.2 → mteb-2.3.3}/mteb.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mteb
-Version: 2.3.2
+Version: 2.3.3
 Summary: Massive Text Embedding Benchmark
 Author-email: MTEB Contributors <niklas@huggingface.co>, Kenneth Enevoldsen <kenneth.enevoldsen@cas.au.dk>, Nouamane Tazi <nouamane@huggingface.co>, Nils Reimers <info@nils-reimers.de>
 Maintainer-email: Kenneth Enevoldsen <kenneth.enevoldsen@cas.au.dk>, Roman Solomatin <risolomatin@gmail.com>, Isaac Chung <chungisaac1217@gmail.com>
@@ -108,7 +108,7 @@ Requires-Dist: qwen_vl_utils>=0.0.14; extra == "eager-embed"
 Dynamic: license-file
 
 <h1 align="center">
-  <img src="docs/images/logos/mteb_logo/dots-icon.png" alt="MTEB" width="28" style="vertical-align: middle; margin-right: 10px;"/> MTEB
+  <img src="https://github.com/embeddings-benchmark/mteb/blob/main/docs/images/logos/mteb_logo/dots-icon.png?raw=true" alt="MTEB" width="28" style="vertical-align: middle; margin-right: 10px;"/> MTEB
 </h1>
 
 <h3 align="center" style="border-bottom: none;">Multimodal toolbox for evaluating embeddings and retrieval systems</h3>
@@ -137,7 +137,7 @@ Dynamic: license-file
 
 
 <h3 align="center">
-    <a href="https://huggingface.co/spaces/mteb/leaderboard"><img style="float: middle; padding: 10px 10px 10px 10px;" width="60" height="55" src="./docs/images/logos/hf_logo.png" /></a>
+    <a href="https://huggingface.co/spaces/mteb/leaderboard"><img style="float: middle; padding: 10px 10px 10px 10px;" width="60" height="55" src="https://github.com/embeddings-benchmark/mteb/blob/main/docs/images/logos/hf_logo.png?raw=true" /></a>
 </h3>
 
 

{mteb-2.3.2 → mteb-2.3.3}/mteb.egg-info/SOURCES.txt

@@ -1469,6 +1469,7 @@ mteb/models/model_implementations/bmretriever_models.py
 mteb/models/model_implementations/cadet_models.py
 mteb/models/model_implementations/cde_models.py
 mteb/models/model_implementations/clip_models.py
+mteb/models/model_implementations/clips_models.py
 mteb/models/model_implementations/codefuse_models.py
 mteb/models/model_implementations/codesage_models.py
 mteb/models/model_implementations/cohere_models.py
@@ -1501,6 +1502,7 @@ mteb/models/model_implementations/jasper_models.py
 mteb/models/model_implementations/jina_clip.py
 mteb/models/model_implementations/jina_models.py
 mteb/models/model_implementations/kalm_models.py
+mteb/models/model_implementations/kennethenevoldsen_models.py
 mteb/models/model_implementations/lens_models.py
 mteb/models/model_implementations/lgai_embedding_models.py
 mteb/models/model_implementations/linq_models.py

{mteb-2.3.2 → mteb-2.3.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mteb"
-version = "2.3.2"
+version = "2.3.3"
 description = "Massive Text Embedding Benchmark"
 readme = "README.md"
 authors = [