mteb 2.7.18__py3-none-any.whl → 2.7.19__py3-none-any.whl

mteb/models/model_implementations/ict_time_and_querit_models.py

@@ -0,0 +1,115 @@
+from __future__ import annotations
+
+from mteb.models.instruct_wrapper import InstructSentenceTransformerModel
+from mteb.models.model_meta import ModelMeta
+from mteb.types import PromptType
+
+
+def instruction_template(
+    instruction: str | dict, prompt_type: PromptType | None = None
+) -> str:
+    """Format instruction for the model."""
+    if isinstance(instruction, dict):
+        instruction = instruction.get(prompt_type.value if prompt_type else "", "")
+    elif prompt_type == PromptType.document:
+        return ""
+
+    if not instruction:
+        return ""
+    return f"Instruct: {instruction}\nQuery:"
+
+
+multilingual_langs = [
+    "deu-Latn",
+    "ita-Latn",
+    "ara-Arab",
+    "fas-Arab",
+    "fra-Latn",
+    "hin-Deva",
+    "spa-Latn",
+    "zho-Hans",
+    "ben-Beng",
+    "eng-Latn",
+    "fin-Latn",
+    "ind-Latn",
+    "jpn-Jpan",
+    "kor-Hang",
+    "rus-Cyrl",
+    "swh-Latn",
+    "tel-Telu",
+    "tha-Thai",
+]
+
+training_data = [
+    "FEVER",
+    "DuRetrieval",
+    "HotpotQA",
+    "MSMARCO",
+    "T2Retrieval",
+    "NQ",
+    "MIRACLRetrieval",
+    "MrTidyRetrieval",
+    "AmazonCounterfactualClassification",
+    "Banking77Classification",
+    "ImdbClassification",
+    "MTOPDomainClassification",
+    "ToxicConversationsClassification",
+    "TweetSentimentExtractionClassification",
+]
+
+boom_4b_instructions = {
+    "AmazonCounterfactualClassification": "Classify a given Amazon customer review text as either counterfactual or not-counterfactual.",
+    "AmazonPolarityClassification": "Classify Amazon reviews into positive or negative sentiment.",
+    "AmazonReviewsClassification": "Classify the given Amazon review into its appropriate rating category.",
+    "Banking77Classification": "Given a online banking query, find the corresponding intents.",
+    "EmotionClassification": "Classify the emotion expressed in the given Twitter message into one of the six emotions: anger, fear, joy, love, sadness, and surprise.",
+    "ImdbClassification": "Classify the sentiment expressed in the given movie review text from the IMDB dataset.",
+    "MassiveIntentClassification": "Given a user utterance as query, find the user intents.",
+    "MassiveScenarioClassification": "Given a user utterance as query, find the user scenarios.",
+    "MTOPDomainClassification": "Classify the intent domain of the given utterance in task-oriented conversation.",
+    "MTOPIntentClassification": "Classify the intent of the given utterance in task-oriented conversation.",
+    "ToxicConversationsClassification": "Classify the given comments as either toxic or not toxic.",
+    "TweetSentimentExtractionClassification": "Classify the sentiment of a given tweet as either positive, negative, or neutral.",
+    "TNews": "Classify the fine-grained category of the given news title.",
+    "ClimateFEVER": "Given a claim about climate change, retrieve documents that support or refute the claim.",
+    "ClimateFEVERHardNegatives": "Given a claim about climate change, retrieve documents that support or refute the claim.",
+    "DBPedia": "Given a query, retrieve relevant entity descriptions from DBPedia.",
+    "FEVER": "Given a claim, retrieve documents that support or refute the claim.",
+    "FEVERHardNegatives": "Given a claim, retrieve documents that support or refute the claim.",
+    "FiQA2018": "Given a financial question, retrieve user replies that best answer the question.",
+    "HotpotQA": "Given a multi-hop question, retrieve documents that can help answer the question.",
+    "HotpotQAHardNegatives": "Given a multi-hop question, retrieve documents that can help answer the question.",
+    "MSMARCO": "Given a web search query, retrieve relevant passages that answer the query.",
+    "NFCorpus": "Given a question, retrieve relevant documents that best answer the question.",
+    "NQ": "Given a question, retrieve Wikipedia passages that answer the question.",
+}
+# How the template actually renders each one at inference time:
+#   instruction_template(boom_4b_instructions["Banking77Classification"], PromptType.query)
+#   -> "Instruct: Given a online banking query, find the corresponding intents.\nQuery:"
+
+boom_4b_v1 = ModelMeta(
+    loader=InstructSentenceTransformerModel,
+    loader_kwargs=dict(
+        instruction_template=instruction_template,
+    ),
+    name="ICT-TIME-and-Querit/BOOM_4B_v1",
+    model_type=["dense"],
+    languages=multilingual_langs,
+    open_weights=True,
+    adapted_from="Qwen/Qwen3-4B",
+    revision="447ab88574d27e67c428acc2b429d7d4580a4ea7",
+    release_date="2026-01-31",
+    n_parameters=4021774336,
+    n_embedding_parameters=None,
+    memory_usage_mb=7671,
+    embed_dim=2560,
+    max_tokens=32768,
+    license="apache-2.0",
+    reference="https://huggingface.co/ICT-TIME-and-Querit/BOOM_4B_v1",
+    similarity_fn_name="cosine",
+    framework=["Sentence Transformers", "PyTorch", "safetensors", "Transformers"],
+    use_instructions=True,
+    public_training_code=None,
+    public_training_data=None,
+    training_datasets=training_data,
+)

mteb/types/_encoder_io.py

@@ -27,7 +27,7 @@ class EncodeKwargs(TypedDict):
 
 
 # --- Output types ---
-Array = NDArray[np.floating | np.integer | np.bool] | torch.Tensor
+Array = NDArray[np.floating | np.integer | np.bool_] | torch.Tensor
 """General array type, can be a numpy array (float, int, or bool) or a torch tensor."""
 
 

{mteb-2.7.18.dist-info → mteb-2.7.19.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mteb
-Version: 2.7.18
+Version: 2.7.19
 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>

{mteb-2.7.18.dist-info → mteb-2.7.19.dist-info}/RECORD

@@ -1539,6 +1539,7 @@ mteb/models/model_implementations/gte_models.py,sha256=-ASkoAuAiVytVtsYMtuKonUf3
 mteb/models/model_implementations/hinvec_models.py,sha256=SYWGFr8XALmM7B9tIHEQnrqq9kZOZIBkW7m7QpzerHI,1756
 mteb/models/model_implementations/human.py,sha256=k7vN6WTcSWyWS9wnluzr6yCOjuMi5LupQnT-4cfzNOk,600
 mteb/models/model_implementations/ibm_granite_models.py,sha256=ipLRDBerTQiL5NaoaDho410Fzy7eNFlF3jB54hGZrwI,8687
+mteb/models/model_implementations/ict_time_and_querit_models.py,sha256=2tR3tLruumZwr5gpyFqott4nznftdOBFOV47-iEl3oI,4993
 mteb/models/model_implementations/inf_models.py,sha256=q_hNNhzMjAxbnJnAT0N6KaNegX_3XZlmz-LXY5C891I,3093
 mteb/models/model_implementations/jasper_models.py,sha256=ourAMx1_L6b2AxX046wQcxDqvYzY1Mx3gaHww0WaMA8,16476
 mteb/models/model_implementations/jina_clip.py,sha256=OF-aC5L8V57-kLdgqyo74S80_q0pxEvc5cyH26Mtwbk,6711
@@ -2641,14 +2642,14 @@ mteb/tasks/zeroshot_classification/eng/sun397.py,sha256=Nls7tXM2Svu008MmAUjt-o_N
 mteb/tasks/zeroshot_classification/eng/ucf101.py,sha256=kwNRYks-_Oe4VE3GyoHIvN-2OJ6zhkwFr76WDNL9ymU,1884
 mteb/tasks/zeroshot_classification/eng/templates/__init__.py,sha256=da1PTClDMl-IBkrSvq6JC1lnS-K_BASzCvxVhNxN5Ls,13
 mteb/types/__init__.py,sha256=O26vXPolPReX7iVUBgUsyCkCo4w8KeLs7uueQDWp3fc,1142
-mteb/types/_encoder_io.py,sha256=V7m_t7ZXm3COJ4SoHP8bcr23WgjFBRCGa9AIaqAX8v4,5939
+mteb/types/_encoder_io.py,sha256=vdIv0_HR1PnJXLPyM4kHmpocT3DvHIbcZP1ue7aU10c,5940
 mteb/types/_metadata.py,sha256=NN-W0S6a5TDV7UkpRx1pyWtGF4TyyCyoPUfHOwdeci8,2290
 mteb/types/_result.py,sha256=UKNokV9pu3G74MGebocU512aU_fFU9I9nPKnrG9Q0iE,1035
 mteb/types/_string_validators.py,sha256=PY-dYq4E8O50VS3bLYdldPWp400fl_WzUjfVSkNWe8U,523
 mteb/types/statistics.py,sha256=gElgSShKBXpfcqaZHhU_d2UHln1CyzUj8FN8KFun_UA,4087
-mteb-2.7.18.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-mteb-2.7.18.dist-info/METADATA,sha256=wIhLzuH5ewCPsURRgO5-DyaJUE6UwGLoZuUfRPBQQzg,14348
-mteb-2.7.18.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-mteb-2.7.18.dist-info/entry_points.txt,sha256=8IJoEJFKoDHmVnNev-qJ9pp4Ln7_1-ma9QsXnzVCzGU,39
-mteb-2.7.18.dist-info/top_level.txt,sha256=OLVIjcQAlWBz0bdmutKlWHLF42FF0hp4uVAg3ZyiG4U,5
-mteb-2.7.18.dist-info/RECORD,,
+mteb-2.7.19.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+mteb-2.7.19.dist-info/METADATA,sha256=y3sFllzuYQsMdkp6mwS6f6bCkQH4hibXb44oEMMCQsY,14348
+mteb-2.7.19.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+mteb-2.7.19.dist-info/entry_points.txt,sha256=8IJoEJFKoDHmVnNev-qJ9pp4Ln7_1-ma9QsXnzVCzGU,39
+mteb-2.7.19.dist-info/top_level.txt,sha256=OLVIjcQAlWBz0bdmutKlWHLF42FF0hp4uVAg3ZyiG4U,5
+mteb-2.7.19.dist-info/RECORD,,