Text Ranking
sentence-transformers
Safetensors
xlm-roberta
cross-encoder
reranker
Generated from Trainer
dataset_size:6763
loss:BinaryCrossEntropyLoss
Eval Results (legacy)
text-embeddings-inference
Instructions to use IoannisKat1/bge_reranker_ft with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use IoannisKat1/bge_reranker_ft with sentence-transformers:
from sentence_transformers import CrossEncoder model = CrossEncoder("IoannisKat1/bge_reranker_ft") query = "Which planet is known as the Red Planet?" passages = [ "Venus is often called Earth's twin because of its similar size and proximity.", "Mars, known for its reddish appearance, is often referred to as the Red Planet.", "Jupiter, the largest planet in our solar system, has a prominent red spot.", "Saturn, famous for its rings, is sometimes mistaken for the Red Planet." ] scores = model.predict([(query, passage) for passage in passages]) print(scores) - Notebooks
- Google Colab
- Kaggle
Finetuned Reranker
Browse files- .gitattributes +1 -0
- README.md +448 -0
- config.json +37 -0
- model.safetensors +3 -0
- sentencepiece.bpe.model +3 -0
- special_tokens_map.json +51 -0
- tokenizer.json +3 -0
- tokenizer_config.json +56 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,448 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- sentence-transformers
|
| 4 |
+
- cross-encoder
|
| 5 |
+
- reranker
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- dataset_size:443
|
| 8 |
+
- loss:BinaryCrossEntropyLoss
|
| 9 |
+
base_model: BAAI/bge-reranker-v2-m3
|
| 10 |
+
pipeline_tag: text-ranking
|
| 11 |
+
library_name: sentence-transformers
|
| 12 |
+
metrics:
|
| 13 |
+
- map
|
| 14 |
+
- mrr@10
|
| 15 |
+
- ndcg@10
|
| 16 |
+
- accuracy
|
| 17 |
+
- accuracy_threshold
|
| 18 |
+
- f1
|
| 19 |
+
- f1_threshold
|
| 20 |
+
- precision
|
| 21 |
+
- recall
|
| 22 |
+
- average_precision
|
| 23 |
+
model-index:
|
| 24 |
+
- name: CrossEncoder based on BAAI/bge-reranker-v2-m3
|
| 25 |
+
results:
|
| 26 |
+
- task:
|
| 27 |
+
type: cross-encoder-reranking
|
| 28 |
+
name: Cross Encoder Reranking
|
| 29 |
+
dataset:
|
| 30 |
+
name: gooaq dev
|
| 31 |
+
type: gooaq-dev
|
| 32 |
+
metrics:
|
| 33 |
+
- type: map
|
| 34 |
+
value: 0.6246
|
| 35 |
+
name: Map
|
| 36 |
+
- type: mrr@10
|
| 37 |
+
value: 0.6246
|
| 38 |
+
name: Mrr@10
|
| 39 |
+
- type: ndcg@10
|
| 40 |
+
value: 0.7177
|
| 41 |
+
name: Ndcg@10
|
| 42 |
+
- task:
|
| 43 |
+
type: cross-encoder-classification
|
| 44 |
+
name: Cross Encoder Classification
|
| 45 |
+
dataset:
|
| 46 |
+
name: cls dev
|
| 47 |
+
type: cls-dev
|
| 48 |
+
metrics:
|
| 49 |
+
- type: accuracy
|
| 50 |
+
value: 0.9523809523809523
|
| 51 |
+
name: Accuracy
|
| 52 |
+
- type: accuracy_threshold
|
| 53 |
+
value: 6.58235585433431e-05
|
| 54 |
+
name: Accuracy Threshold
|
| 55 |
+
- type: f1
|
| 56 |
+
value: 0.975609756097561
|
| 57 |
+
name: F1
|
| 58 |
+
- type: f1_threshold
|
| 59 |
+
value: 6.58235585433431e-05
|
| 60 |
+
name: F1 Threshold
|
| 61 |
+
- type: precision
|
| 62 |
+
value: 1.0
|
| 63 |
+
name: Precision
|
| 64 |
+
- type: recall
|
| 65 |
+
value: 0.9523809523809523
|
| 66 |
+
name: Recall
|
| 67 |
+
- type: average_precision
|
| 68 |
+
value: 1.0000000000000002
|
| 69 |
+
name: Average Precision
|
| 70 |
+
---
|
| 71 |
+
|
| 72 |
+
# CrossEncoder based on BAAI/bge-reranker-v2-m3
|
| 73 |
+
|
| 74 |
+
This is a [Cross Encoder](https://www.sbert.net/docs/cross_encoder/usage/usage.html) model finetuned from [BAAI/bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3) using the [sentence-transformers](https://www.SBERT.net) library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.
|
| 75 |
+
|
| 76 |
+
## Model Details
|
| 77 |
+
|
| 78 |
+
### Model Description
|
| 79 |
+
- **Model Type:** Cross Encoder
|
| 80 |
+
- **Base model:** [BAAI/bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3) <!-- at revision 953dc6f6f85a1b2dbfca4c34a2796e7dde08d41e -->
|
| 81 |
+
- **Maximum Sequence Length:** 512 tokens
|
| 82 |
+
- **Number of Output Labels:** 1 label
|
| 83 |
+
<!-- - **Training Dataset:** Unknown -->
|
| 84 |
+
<!-- - **Language:** Unknown -->
|
| 85 |
+
<!-- - **License:** Unknown -->
|
| 86 |
+
|
| 87 |
+
### Model Sources
|
| 88 |
+
|
| 89 |
+
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
|
| 90 |
+
- **Documentation:** [Cross Encoder Documentation](https://www.sbert.net/docs/cross_encoder/usage/usage.html)
|
| 91 |
+
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
|
| 92 |
+
- **Hugging Face:** [Cross Encoders on Hugging Face](https://huggingface.co/models?library=sentence-transformers&other=cross-encoder)
|
| 93 |
+
|
| 94 |
+
## Usage
|
| 95 |
+
|
| 96 |
+
### Direct Usage (Sentence Transformers)
|
| 97 |
+
|
| 98 |
+
First install the Sentence Transformers library:
|
| 99 |
+
|
| 100 |
+
```bash
|
| 101 |
+
pip install -U sentence-transformers
|
| 102 |
+
```
|
| 103 |
+
|
| 104 |
+
Then you can load this model and run inference.
|
| 105 |
+
```python
|
| 106 |
+
from sentence_transformers import CrossEncoder
|
| 107 |
+
|
| 108 |
+
# Download from the 🤗 Hub
|
| 109 |
+
model = CrossEncoder("cross_encoder_model_id")
|
| 110 |
+
# Get scores for pairs of texts
|
| 111 |
+
pairs = [
|
| 112 |
+
['What does dolus refer to?', 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision, it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without requiring that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and performs an act, omission, or acquiescence; and\n\nc) Damage to another’s property, according to civil law, which must be causally connected to the perpetrator’s deceptive acts or omissions. It is not required that the deceived person and the person who suffered the loss be the same.\n\nThe term “facts,” within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts relating to the present or the past, in such a way as to create the impression of future fulfillment, based on a false present situation fabricated by the perpetrator—who has already made the decision not to fulfill their obligation—then the crime of fraud is established.\n\nThe term “property” denotes the totality of a person’s economic assets possessing monetary value, while damage to property refers to its reduction—specifically, the difference between the property’s monetary value before the disposition caused by the fraudulent conduct and its value afterward. Property damage exists even if the victim has an active claim for its restitution.\n\nThe time of commission of fraud is considered to be the moment when the perpetrator acted and completed the deceptive conduct, that is, when they made the false representations which deceived the victim or a third party. Any later time at which the victim’s financial loss occurred—thus completing the fraud—or the time when the harmful act or omission of the deceived person took place, is irrelevant.\n\nThe reference to multiple modes of commission of fraud (i.e., both the misrepresentation of false facts and the concealment of true ones) may create ambiguity and contradiction, unless it is made clear from the overall findings that the offense was committed in one particular manner, and that the reference to the other merely serves to define the intent (mens rea) of the perpetrator—specifically, that the representations were false.\n\nFurthermore, a conviction must contain the specific and well-reasoned justification required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of Criminal Procedure. The absence of such reasoning constitutes grounds for cassation (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when the judgment does not set out, with clarity, completeness, and consistency, the factual circumstances established by the evidence, upon which the court based its findings regarding the objective and subjective elements of the offense, the evidence supporting those findings, and the legal reasoning through which those facts were subsumed under the applicable substantive criminal provision.\n\nFor the existence of such reasoning, the explanatory and operative parts of the decision may complement each other, as they form a single, unified whole.\n\nThe existence of intent (dolus) does not generally need to be specially justified, since it is inherent in the will to bring about the factual circumstances constituting the objective elements of the offense, and it is presumed from their realization in each particular case—unless the law requires additional elements for criminal liability, such as the act being committed with knowledge of a specific circumstance (direct intent) or with the pursuit of a further purpose, i.e., the achievement of an additional result (offenses requiring a special subjective element).\n\nFurthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure, a misapplication of substantive criminal law also constitutes grounds for cassation. Such misapplication occurs when the trial court incorrectly applies the law to the facts it has found to be true, or when the violation occurs indirectly, namely when the reasoning of the judgment—comprising the combination of its factual and operative parts and relating to the elements and identity of the offense—contains ambiguities, contradictions, or logical gaps, rendering it impossible to verify, on appeal, whether the law was applied correctly. In such cases, the judgment lacks a lawful basis.'],
|
| 113 |
+
['What does dolus refer to?', 'According to Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom these provisions, it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves or another; and\n\nc) Damage to another’s property, as defined under civil law, which must be causally connected to the perpetrator’s deceptive acts.\n\nFrom the above provisions, it is deduced that the crime of fraud is established both objectively and subjectively through the knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true ones, by which another person is deceived and, as a result, performs an act, omission, or acquiescence involving a disposition of property that directly and necessarily causes financial damage to the deceived person or another, with the intent that the perpetrator or another gain an unlawful benefit. It is irrelevant whether this intended benefit was ultimately achieved.\n\nThe term “facts,” within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those expected to occur in the future, such as mere promises or contractual obligations. The false fact must have existed in the past or must be a present circumstance at the time it is asserted, and cannot relate to the future.\n\nHowever, when future circumstances—that is, promises or contractual obligations—are accompanied by false assurances and representations of other false facts referring to the present or past, in such a way as to create the impression of future fulfillment, based on a false present situation or supposed ability of the perpetrator, who had already made the decision not to fulfill their obligation, then the crime of fraud is established.'],
|
| 114 |
+
['What does dolus refer to?', 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without it being necessary that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence that is detrimental to themselves or another; and\n\nc) Damage to another person’s property, as defined under civil law, which must be causally linked to the deceptive acts or omissions of the perpetrator. It is not required that the person deceived and the person who suffered the damage be the same individual.\n\nThe term “facts”, within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts referring to the present or the past, in such a manner as to create the impression of future fulfillment based on a false present situation fabricated by the perpetrator, who has already formed the decision not to fulfill their obligation, the crime of fraud is established.\n\nThe term “property” refers to the totality of a person’s economic assets that possess monetary value, while damage to property means its reduction—specifically, the difference between the monetary value the property had before the disposition caused by the fraudulent conduct and the value remaining after it. Property damage exists even if the victim possesses an active claim for restitution.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their fraudulent conduct, namely when they made the false representations that deceived the victim or a third party. Any subsequent moment at which the victim’s damage actually occurred—thereby completing the fraud—or the time when the victim carried out the harmful act or omission, is irrelevant.'],
|
| 115 |
+
['What does dolus refer to?', 'According to Article 386 paragraph 1 of the Greek Penal Code, the crime of fraud is established both objectively and subjectively when a person knowingly presents false facts as true, or unlawfully conceals or suppresses true facts, thereby deceiving another person who, as a result of this deception, performs an act, omission, or acquiescence that involves a disposition of property, which directly and necessarily causes financial harm to the deceived person or another, with the intent of obtaining an unlawful benefit for themselves or another. It is irrelevant whether or not this intended benefit was actually achieved.\n\nFurthermore, within the meaning of the law, “facts” also include those referring to future events and promises, when they are accompanied by false assurances and representations of other false facts relating to the past or present, in such a manner as to create the impression of future fulfillment, based on the false situation presented by the perpetrator, who has already made the decision not to fulfill their obligation.'],
|
| 116 |
+
['What does dolus refer to?', 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another person’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision, it follows that for the crime of fraud to be established, the following elements are required:\n\na) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, regardless of whether this benefit was actually realized;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which, as a causal factor, someone is deceived and acts in a way that is detrimental to themselves or another (by an act, omission, or acquiescence); and\n\nc) Damage to another’s property, in the sense recognized by civil law, which must be causally linked to the fraudulent conduct (the deceptive act or omission of the perpetrator) and to the resulting deception of the person who made the property disposition. It is not required that the person deceived be the same person who suffered the damage.\n\nProperty damage exists when there is a reduction or deterioration in the victim’s assets, even if the victim has an active claim to restitution. However, as an element of the objective aspect of the crime of fraud, the damage must be the direct, necessary, and exclusive result of the property disposition—namely, the act, omission, or acquiescence performed by the person deceived by the perpetrator’s fraudulent conduct.\n\nThere must therefore be a causal connection between the perpetrator’s deceptive behavior and the deception it caused, as well as between this deception and the resulting property damage, which must be the direct, necessary, and exclusive outcome of the deception and of the act, omission, or acquiescence of the deceived person.\n\nThe term “facts” refers to real circumstances relating to the past or present, and not to those expected to occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts relating to the present or the past, in such a way as to create the impression of future fulfillment, based on the false present situation presented by a perpetrator who has already made the decision not to fulfill their obligation, then the crime of fraud is established.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their deceptive conduct—that is, when they made the false representations that deceived the victim or a third party. Any later time at which the victim’s financial loss actually occurred—thus completing the fraud—or the time when the deceived person performed the harmful act or omission, is irrelevant.'],
|
| 117 |
+
]
|
| 118 |
+
scores = model.predict(pairs)
|
| 119 |
+
print(scores.shape)
|
| 120 |
+
# (5,)
|
| 121 |
+
|
| 122 |
+
# Or rank different texts based on similarity to a single text
|
| 123 |
+
ranks = model.rank(
|
| 124 |
+
'What does dolus refer to?',
|
| 125 |
+
[
|
| 126 |
+
'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision, it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without requiring that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and performs an act, omission, or acquiescence; and\n\nc) Damage to another’s property, according to civil law, which must be causally connected to the perpetrator’s deceptive acts or omissions. It is not required that the deceived person and the person who suffered the loss be the same.\n\nThe term “facts,” within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts relating to the present or the past, in such a way as to create the impression of future fulfillment, based on a false present situation fabricated by the perpetrator—who has already made the decision not to fulfill their obligation—then the crime of fraud is established.\n\nThe term “property” denotes the totality of a person’s economic assets possessing monetary value, while damage to property refers to its reduction—specifically, the difference between the property’s monetary value before the disposition caused by the fraudulent conduct and its value afterward. Property damage exists even if the victim has an active claim for its restitution.\n\nThe time of commission of fraud is considered to be the moment when the perpetrator acted and completed the deceptive conduct, that is, when they made the false representations which deceived the victim or a third party. Any later time at which the victim’s financial loss occurred—thus completing the fraud—or the time when the harmful act or omission of the deceived person took place, is irrelevant.\n\nThe reference to multiple modes of commission of fraud (i.e., both the misrepresentation of false facts and the concealment of true ones) may create ambiguity and contradiction, unless it is made clear from the overall findings that the offense was committed in one particular manner, and that the reference to the other merely serves to define the intent (mens rea) of the perpetrator—specifically, that the representations were false.\n\nFurthermore, a conviction must contain the specific and well-reasoned justification required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of Criminal Procedure. The absence of such reasoning constitutes grounds for cassation (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when the judgment does not set out, with clarity, completeness, and consistency, the factual circumstances established by the evidence, upon which the court based its findings regarding the objective and subjective elements of the offense, the evidence supporting those findings, and the legal reasoning through which those facts were subsumed under the applicable substantive criminal provision.\n\nFor the existence of such reasoning, the explanatory and operative parts of the decision may complement each other, as they form a single, unified whole.\n\nThe existence of intent (dolus) does not generally need to be specially justified, since it is inherent in the will to bring about the factual circumstances constituting the objective elements of the offense, and it is presumed from their realization in each particular case—unless the law requires additional elements for criminal liability, such as the act being committed with knowledge of a specific circumstance (direct intent) or with the pursuit of a further purpose, i.e., the achievement of an additional result (offenses requiring a special subjective element).\n\nFurthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure, a misapplication of substantive criminal law also constitutes grounds for cassation. Such misapplication occurs when the trial court incorrectly applies the law to the facts it has found to be true, or when the violation occurs indirectly, namely when the reasoning of the judgment—comprising the combination of its factual and operative parts and relating to the elements and identity of the offense—contains ambiguities, contradictions, or logical gaps, rendering it impossible to verify, on appeal, whether the law was applied correctly. In such cases, the judgment lacks a lawful basis.',
|
| 127 |
+
'According to Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom these provisions, it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves or another; and\n\nc) Damage to another’s property, as defined under civil law, which must be causally connected to the perpetrator’s deceptive acts.\n\nFrom the above provisions, it is deduced that the crime of fraud is established both objectively and subjectively through the knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true ones, by which another person is deceived and, as a result, performs an act, omission, or acquiescence involving a disposition of property that directly and necessarily causes financial damage to the deceived person or another, with the intent that the perpetrator or another gain an unlawful benefit. It is irrelevant whether this intended benefit was ultimately achieved.\n\nThe term “facts,” within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those expected to occur in the future, such as mere promises or contractual obligations. The false fact must have existed in the past or must be a present circumstance at the time it is asserted, and cannot relate to the future.\n\nHowever, when future circumstances—that is, promises or contractual obligations—are accompanied by false assurances and representations of other false facts referring to the present or past, in such a way as to create the impression of future fulfillment, based on a false present situation or supposed ability of the perpetrator, who had already made the decision not to fulfill their obligation, then the crime of fraud is established.',
|
| 128 |
+
'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without it being necessary that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence that is detrimental to themselves or another; and\n\nc) Damage to another person’s property, as defined under civil law, which must be causally linked to the deceptive acts or omissions of the perpetrator. It is not required that the person deceived and the person who suffered the damage be the same individual.\n\nThe term “facts”, within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts referring to the present or the past, in such a manner as to create the impression of future fulfillment based on a false present situation fabricated by the perpetrator, who has already formed the decision not to fulfill their obligation, the crime of fraud is established.\n\nThe term “property” refers to the totality of a person’s economic assets that possess monetary value, while damage to property means its reduction—specifically, the difference between the monetary value the property had before the disposition caused by the fraudulent conduct and the value remaining after it. Property damage exists even if the victim possesses an active claim for restitution.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their fraudulent conduct, namely when they made the false representations that deceived the victim or a third party. Any subsequent moment at which the victim’s damage actually occurred—thereby completing the fraud—or the time when the victim carried out the harmful act or omission, is irrelevant.',
|
| 129 |
+
'According to Article 386 paragraph 1 of the Greek Penal Code, the crime of fraud is established both objectively and subjectively when a person knowingly presents false facts as true, or unlawfully conceals or suppresses true facts, thereby deceiving another person who, as a result of this deception, performs an act, omission, or acquiescence that involves a disposition of property, which directly and necessarily causes financial harm to the deceived person or another, with the intent of obtaining an unlawful benefit for themselves or another. It is irrelevant whether or not this intended benefit was actually achieved.\n\nFurthermore, within the meaning of the law, “facts” also include those referring to future events and promises, when they are accompanied by false assurances and representations of other false facts relating to the past or present, in such a manner as to create the impression of future fulfillment, based on the false situation presented by the perpetrator, who has already made the decision not to fulfill their obligation.',
|
| 130 |
+
'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another person’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision, it follows that for the crime of fraud to be established, the following elements are required:\n\na) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, regardless of whether this benefit was actually realized;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which, as a causal factor, someone is deceived and acts in a way that is detrimental to themselves or another (by an act, omission, or acquiescence); and\n\nc) Damage to another’s property, in the sense recognized by civil law, which must be causally linked to the fraudulent conduct (the deceptive act or omission of the perpetrator) and to the resulting deception of the person who made the property disposition. It is not required that the person deceived be the same person who suffered the damage.\n\nProperty damage exists when there is a reduction or deterioration in the victim’s assets, even if the victim has an active claim to restitution. However, as an element of the objective aspect of the crime of fraud, the damage must be the direct, necessary, and exclusive result of the property disposition—namely, the act, omission, or acquiescence performed by the person deceived by the perpetrator’s fraudulent conduct.\n\nThere must therefore be a causal connection between the perpetrator’s deceptive behavior and the deception it caused, as well as between this deception and the resulting property damage, which must be the direct, necessary, and exclusive outcome of the deception and of the act, omission, or acquiescence of the deceived person.\n\nThe term “facts” refers to real circumstances relating to the past or present, and not to those expected to occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts relating to the present or the past, in such a way as to create the impression of future fulfillment, based on the false present situation presented by a perpetrator who has already made the decision not to fulfill their obligation, then the crime of fraud is established.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their deceptive conduct—that is, when they made the false representations that deceived the victim or a third party. Any later time at which the victim’s financial loss actually occurred—thus completing the fraud—or the time when the deceived person performed the harmful act or omission, is irrelevant.',
|
| 131 |
+
]
|
| 132 |
+
)
|
| 133 |
+
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]
|
| 134 |
+
```
|
| 135 |
+
|
| 136 |
+
<!--
|
| 137 |
+
### Direct Usage (Transformers)
|
| 138 |
+
|
| 139 |
+
<details><summary>Click to see the direct usage in Transformers</summary>
|
| 140 |
+
|
| 141 |
+
</details>
|
| 142 |
+
-->
|
| 143 |
+
|
| 144 |
+
<!--
|
| 145 |
+
### Downstream Usage (Sentence Transformers)
|
| 146 |
+
|
| 147 |
+
You can finetune this model on your own dataset.
|
| 148 |
+
|
| 149 |
+
<details><summary>Click to expand</summary>
|
| 150 |
+
|
| 151 |
+
</details>
|
| 152 |
+
-->
|
| 153 |
+
|
| 154 |
+
<!--
|
| 155 |
+
### Out-of-Scope Use
|
| 156 |
+
|
| 157 |
+
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
|
| 158 |
+
-->
|
| 159 |
+
|
| 160 |
+
## Evaluation
|
| 161 |
+
|
| 162 |
+
### Metrics
|
| 163 |
+
|
| 164 |
+
#### Cross Encoder Reranking
|
| 165 |
+
|
| 166 |
+
* Dataset: `gooaq-dev`
|
| 167 |
+
* Evaluated with [<code>CrossEncoderRerankingEvaluator</code>](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderRerankingEvaluator) with these parameters:
|
| 168 |
+
```json
|
| 169 |
+
{
|
| 170 |
+
"at_k": 10,
|
| 171 |
+
"always_rerank_positives": false
|
| 172 |
+
}
|
| 173 |
+
```
|
| 174 |
+
|
| 175 |
+
| Metric | Value |
|
| 176 |
+
|:------------|:---------------------|
|
| 177 |
+
| map | 0.6246 (-0.0024) |
|
| 178 |
+
| mrr@10 | 0.6246 (-0.0024) |
|
| 179 |
+
| **ndcg@10** | **0.7177 (-0.0021)** |
|
| 180 |
+
|
| 181 |
+
#### Cross Encoder Classification
|
| 182 |
+
|
| 183 |
+
* Dataset: `cls-dev`
|
| 184 |
+
* Evaluated with [<code>CrossEncoderClassificationEvaluator</code>](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderClassificationEvaluator)
|
| 185 |
+
|
| 186 |
+
| Metric | Value |
|
| 187 |
+
|:----------------------|:--------|
|
| 188 |
+
| accuracy | 0.9524 |
|
| 189 |
+
| accuracy_threshold | 0.0001 |
|
| 190 |
+
| f1 | 0.9756 |
|
| 191 |
+
| f1_threshold | 0.0001 |
|
| 192 |
+
| precision | 1.0 |
|
| 193 |
+
| recall | 0.9524 |
|
| 194 |
+
| **average_precision** | **1.0** |
|
| 195 |
+
|
| 196 |
+
<!--
|
| 197 |
+
## Bias, Risks and Limitations
|
| 198 |
+
|
| 199 |
+
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
|
| 200 |
+
-->
|
| 201 |
+
|
| 202 |
+
<!--
|
| 203 |
+
### Recommendations
|
| 204 |
+
|
| 205 |
+
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
|
| 206 |
+
-->
|
| 207 |
+
|
| 208 |
+
## Training Details
|
| 209 |
+
|
| 210 |
+
### Training Dataset
|
| 211 |
+
|
| 212 |
+
#### Unnamed Dataset
|
| 213 |
+
|
| 214 |
+
* Size: 443 training samples
|
| 215 |
+
* Columns: <code>query</code>, <code>response</code>, and <code>label</code>
|
| 216 |
+
* Approximate statistics based on the first 443 samples:
|
| 217 |
+
| | query | response | label |
|
| 218 |
+
|:--------|:------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------|:------------------------------------------------|
|
| 219 |
+
| type | string | string | int |
|
| 220 |
+
| details | <ul><li>min: 25 characters</li><li>mean: 74.23 characters</li><li>max: 167 characters</li></ul> | <ul><li>min: 294 characters</li><li>mean: 1983.03 characters</li><li>max: 5250 characters</li></ul> | <ul><li>0: ~81.49%</li><li>1: ~18.51%</li></ul> |
|
| 221 |
+
* Samples:
|
| 222 |
+
| query | response | label |
|
| 223 |
+
|:---------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
|
| 224 |
+
| <code>What does dolus refer to?</code> | <code>According to the provision of Article 386 paragraph 1 of the Greek Penal Code,<br><br>"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."<br><br>From this provision, it follows that, for the crime of fraud to be established, the following elements are required:<br><br>a) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without requiring that the benefit actually materialize;<br><br>b) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is dece...</code> | <code>1</code> |
|
| 225 |
+
| <code>What does dolus refer to?</code> | <code>According to Article 386 paragraph 1 of the Greek Penal Code,<br><br>"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."<br><br>From these provisions, it follows that, for the crime of fraud to be established, the following elements are required:<br><br>a) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit;<br><br>b) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence detrimental to th...</code> | <code>0</code> |
|
| 226 |
+
| <code>What does dolus refer to?</code> | <code>According to the provision of Article 386 paragraph 1 of the Greek Penal Code,<br><br>"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."<br><br>From this provision it follows that, for the crime of fraud to be established, the following elements are required:<br><br>a) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without it being necessary that the benefit actually materialize;<br><br>b) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone...</code> | <code>0</code> |
|
| 227 |
+
* Loss: [<code>BinaryCrossEntropyLoss</code>](https://sbert.net/docs/package_reference/cross_encoder/losses.html#binarycrossentropyloss) with these parameters:
|
| 228 |
+
```json
|
| 229 |
+
{
|
| 230 |
+
"activation_fn": "torch.nn.modules.linear.Identity",
|
| 231 |
+
"pos_weight": null
|
| 232 |
+
}
|
| 233 |
+
```
|
| 234 |
+
|
| 235 |
+
### Training Hyperparameters
|
| 236 |
+
#### Non-Default Hyperparameters
|
| 237 |
+
|
| 238 |
+
- `eval_strategy`: steps
|
| 239 |
+
- `per_device_train_batch_size`: 16
|
| 240 |
+
- `per_device_eval_batch_size`: 16
|
| 241 |
+
- `learning_rate`: 2e-05
|
| 242 |
+
- `num_train_epochs`: 20
|
| 243 |
+
- `warmup_ratio`: 0.1
|
| 244 |
+
- `seed`: 12
|
| 245 |
+
- `dataloader_num_workers`: 4
|
| 246 |
+
- `load_best_model_at_end`: True
|
| 247 |
+
|
| 248 |
+
#### All Hyperparameters
|
| 249 |
+
<details><summary>Click to expand</summary>
|
| 250 |
+
|
| 251 |
+
- `overwrite_output_dir`: False
|
| 252 |
+
- `do_predict`: False
|
| 253 |
+
- `eval_strategy`: steps
|
| 254 |
+
- `prediction_loss_only`: True
|
| 255 |
+
- `per_device_train_batch_size`: 16
|
| 256 |
+
- `per_device_eval_batch_size`: 16
|
| 257 |
+
- `per_gpu_train_batch_size`: None
|
| 258 |
+
- `per_gpu_eval_batch_size`: None
|
| 259 |
+
- `gradient_accumulation_steps`: 1
|
| 260 |
+
- `eval_accumulation_steps`: None
|
| 261 |
+
- `torch_empty_cache_steps`: None
|
| 262 |
+
- `learning_rate`: 2e-05
|
| 263 |
+
- `weight_decay`: 0.0
|
| 264 |
+
- `adam_beta1`: 0.9
|
| 265 |
+
- `adam_beta2`: 0.999
|
| 266 |
+
- `adam_epsilon`: 1e-08
|
| 267 |
+
- `max_grad_norm`: 1.0
|
| 268 |
+
- `num_train_epochs`: 20
|
| 269 |
+
- `max_steps`: -1
|
| 270 |
+
- `lr_scheduler_type`: linear
|
| 271 |
+
- `lr_scheduler_kwargs`: {}
|
| 272 |
+
- `warmup_ratio`: 0.1
|
| 273 |
+
- `warmup_steps`: 0
|
| 274 |
+
- `log_level`: passive
|
| 275 |
+
- `log_level_replica`: warning
|
| 276 |
+
- `log_on_each_node`: True
|
| 277 |
+
- `logging_nan_inf_filter`: True
|
| 278 |
+
- `save_safetensors`: True
|
| 279 |
+
- `save_on_each_node`: False
|
| 280 |
+
- `save_only_model`: False
|
| 281 |
+
- `restore_callback_states_from_checkpoint`: False
|
| 282 |
+
- `no_cuda`: False
|
| 283 |
+
- `use_cpu`: False
|
| 284 |
+
- `use_mps_device`: False
|
| 285 |
+
- `seed`: 12
|
| 286 |
+
- `data_seed`: None
|
| 287 |
+
- `jit_mode_eval`: False
|
| 288 |
+
- `use_ipex`: False
|
| 289 |
+
- `bf16`: False
|
| 290 |
+
- `fp16`: False
|
| 291 |
+
- `fp16_opt_level`: O1
|
| 292 |
+
- `half_precision_backend`: auto
|
| 293 |
+
- `bf16_full_eval`: False
|
| 294 |
+
- `fp16_full_eval`: False
|
| 295 |
+
- `tf32`: None
|
| 296 |
+
- `local_rank`: 0
|
| 297 |
+
- `ddp_backend`: None
|
| 298 |
+
- `tpu_num_cores`: None
|
| 299 |
+
- `tpu_metrics_debug`: False
|
| 300 |
+
- `debug`: []
|
| 301 |
+
- `dataloader_drop_last`: False
|
| 302 |
+
- `dataloader_num_workers`: 4
|
| 303 |
+
- `dataloader_prefetch_factor`: None
|
| 304 |
+
- `past_index`: -1
|
| 305 |
+
- `disable_tqdm`: False
|
| 306 |
+
- `remove_unused_columns`: True
|
| 307 |
+
- `label_names`: None
|
| 308 |
+
- `load_best_model_at_end`: True
|
| 309 |
+
- `ignore_data_skip`: False
|
| 310 |
+
- `fsdp`: []
|
| 311 |
+
- `fsdp_min_num_params`: 0
|
| 312 |
+
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
|
| 313 |
+
- `tp_size`: 0
|
| 314 |
+
- `fsdp_transformer_layer_cls_to_wrap`: None
|
| 315 |
+
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
|
| 316 |
+
- `deepspeed`: None
|
| 317 |
+
- `label_smoothing_factor`: 0.0
|
| 318 |
+
- `optim`: adamw_torch
|
| 319 |
+
- `optim_args`: None
|
| 320 |
+
- `adafactor`: False
|
| 321 |
+
- `group_by_length`: False
|
| 322 |
+
- `length_column_name`: length
|
| 323 |
+
- `ddp_find_unused_parameters`: None
|
| 324 |
+
- `ddp_bucket_cap_mb`: None
|
| 325 |
+
- `ddp_broadcast_buffers`: False
|
| 326 |
+
- `dataloader_pin_memory`: True
|
| 327 |
+
- `dataloader_persistent_workers`: False
|
| 328 |
+
- `skip_memory_metrics`: True
|
| 329 |
+
- `use_legacy_prediction_loop`: False
|
| 330 |
+
- `push_to_hub`: False
|
| 331 |
+
- `resume_from_checkpoint`: None
|
| 332 |
+
- `hub_model_id`: None
|
| 333 |
+
- `hub_strategy`: every_save
|
| 334 |
+
- `hub_private_repo`: None
|
| 335 |
+
- `hub_always_push`: False
|
| 336 |
+
- `gradient_checkpointing`: False
|
| 337 |
+
- `gradient_checkpointing_kwargs`: None
|
| 338 |
+
- `include_inputs_for_metrics`: False
|
| 339 |
+
- `include_for_metrics`: []
|
| 340 |
+
- `eval_do_concat_batches`: True
|
| 341 |
+
- `fp16_backend`: auto
|
| 342 |
+
- `push_to_hub_model_id`: None
|
| 343 |
+
- `push_to_hub_organization`: None
|
| 344 |
+
- `mp_parameters`:
|
| 345 |
+
- `auto_find_batch_size`: False
|
| 346 |
+
- `full_determinism`: False
|
| 347 |
+
- `torchdynamo`: None
|
| 348 |
+
- `ray_scope`: last
|
| 349 |
+
- `ddp_timeout`: 1800
|
| 350 |
+
- `torch_compile`: False
|
| 351 |
+
- `torch_compile_backend`: None
|
| 352 |
+
- `torch_compile_mode`: None
|
| 353 |
+
- `include_tokens_per_second`: False
|
| 354 |
+
- `include_num_input_tokens_seen`: False
|
| 355 |
+
- `neftune_noise_alpha`: None
|
| 356 |
+
- `optim_target_modules`: None
|
| 357 |
+
- `batch_eval_metrics`: False
|
| 358 |
+
- `eval_on_start`: False
|
| 359 |
+
- `use_liger_kernel`: False
|
| 360 |
+
- `eval_use_gather_object`: False
|
| 361 |
+
- `average_tokens_across_devices`: False
|
| 362 |
+
- `prompts`: None
|
| 363 |
+
- `batch_sampler`: batch_sampler
|
| 364 |
+
- `multi_dataset_batch_sampler`: proportional
|
| 365 |
+
- `router_mapping`: {}
|
| 366 |
+
- `learning_rate_mapping`: {}
|
| 367 |
+
|
| 368 |
+
</details>
|
| 369 |
+
|
| 370 |
+
### Training Logs
|
| 371 |
+
| Epoch | Step | Training Loss | gooaq-dev_ndcg@10 | cls-dev_average_precision |
|
| 372 |
+
|:-------:|:----:|:-------------:|:-----------------:|:-------------------------:|
|
| 373 |
+
| -1 | -1 | - | 0.8520 (+0.1323) | - |
|
| 374 |
+
| 0.0357 | 1 | 0.9641 | - | - |
|
| 375 |
+
| 0.7143 | 20 | 0.5838 | - | - |
|
| 376 |
+
| 1.4286 | 40 | 0.3681 | - | - |
|
| 377 |
+
| 2.1429 | 60 | 0.3216 | - | - |
|
| 378 |
+
| 2.8571 | 80 | 0.3218 | - | - |
|
| 379 |
+
| 3.5714 | 100 | 0.2768 | - | - |
|
| 380 |
+
| 4.2857 | 120 | 0.2571 | - | - |
|
| 381 |
+
| 5.0 | 140 | 0.2517 | - | - |
|
| 382 |
+
| 5.7143 | 160 | 0.214 | - | - |
|
| 383 |
+
| 6.4286 | 180 | 0.1952 | - | - |
|
| 384 |
+
| 7.1429 | 200 | 0.1884 | - | - |
|
| 385 |
+
| 7.8571 | 220 | 0.1793 | - | - |
|
| 386 |
+
| 8.5714 | 240 | 0.1328 | - | - |
|
| 387 |
+
| 9.2857 | 260 | 0.1964 | - | - |
|
| 388 |
+
| 10.0 | 280 | 0.1444 | - | - |
|
| 389 |
+
| 10.7143 | 300 | 0.1408 | - | - |
|
| 390 |
+
| 11.4286 | 320 | 0.1308 | - | - |
|
| 391 |
+
| 12.1429 | 340 | 0.1147 | - | - |
|
| 392 |
+
| 12.8571 | 360 | 0.1495 | - | - |
|
| 393 |
+
| 13.5714 | 380 | 0.0991 | - | - |
|
| 394 |
+
| 14.2857 | 400 | 0.168 | - | - |
|
| 395 |
+
| 15.0 | 420 | 0.0972 | - | - |
|
| 396 |
+
| 15.7143 | 440 | 0.0859 | - | - |
|
| 397 |
+
| 16.4286 | 460 | 0.1162 | - | - |
|
| 398 |
+
| 17.1429 | 480 | 0.0693 | - | - |
|
| 399 |
+
| 17.8571 | 500 | 0.0836 | - | - |
|
| 400 |
+
| 18.5714 | 520 | 0.056 | - | - |
|
| 401 |
+
| 19.2857 | 540 | 0.0727 | - | - |
|
| 402 |
+
| 20.0 | 560 | 0.0715 | - | - |
|
| 403 |
+
| -1 | -1 | - | 0.7177 (-0.0021) | 1.0000 |
|
| 404 |
+
|
| 405 |
+
|
| 406 |
+
### Framework Versions
|
| 407 |
+
- Python: 3.12.12
|
| 408 |
+
- Sentence Transformers: 5.1.1
|
| 409 |
+
- Transformers: 4.51.3
|
| 410 |
+
- PyTorch: 2.8.0+cu126
|
| 411 |
+
- Accelerate: 1.11.0
|
| 412 |
+
- Datasets: 4.0.0
|
| 413 |
+
- Tokenizers: 0.21.4
|
| 414 |
+
|
| 415 |
+
## Citation
|
| 416 |
+
|
| 417 |
+
### BibTeX
|
| 418 |
+
|
| 419 |
+
#### Sentence Transformers
|
| 420 |
+
```bibtex
|
| 421 |
+
@inproceedings{reimers-2019-sentence-bert,
|
| 422 |
+
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
|
| 423 |
+
author = "Reimers, Nils and Gurevych, Iryna",
|
| 424 |
+
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
|
| 425 |
+
month = "11",
|
| 426 |
+
year = "2019",
|
| 427 |
+
publisher = "Association for Computational Linguistics",
|
| 428 |
+
url = "https://arxiv.org/abs/1908.10084",
|
| 429 |
+
}
|
| 430 |
+
```
|
| 431 |
+
|
| 432 |
+
<!--
|
| 433 |
+
## Glossary
|
| 434 |
+
|
| 435 |
+
*Clearly define terms in order to be accessible across audiences.*
|
| 436 |
+
-->
|
| 437 |
+
|
| 438 |
+
<!--
|
| 439 |
+
## Model Card Authors
|
| 440 |
+
|
| 441 |
+
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
|
| 442 |
+
-->
|
| 443 |
+
|
| 444 |
+
<!--
|
| 445 |
+
## Model Card Contact
|
| 446 |
+
|
| 447 |
+
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
|
| 448 |
+
-->
|
config.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"XLMRobertaForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 4096,
|
| 17 |
+
"label2id": {
|
| 18 |
+
"LABEL_0": 0
|
| 19 |
+
},
|
| 20 |
+
"layer_norm_eps": 1e-05,
|
| 21 |
+
"max_position_embeddings": 8194,
|
| 22 |
+
"model_type": "xlm-roberta",
|
| 23 |
+
"num_attention_heads": 16,
|
| 24 |
+
"num_hidden_layers": 24,
|
| 25 |
+
"output_past": true,
|
| 26 |
+
"pad_token_id": 1,
|
| 27 |
+
"position_embedding_type": "absolute",
|
| 28 |
+
"sentence_transformers": {
|
| 29 |
+
"activation_fn": "torch.nn.modules.activation.Sigmoid",
|
| 30 |
+
"version": "5.1.1"
|
| 31 |
+
},
|
| 32 |
+
"torch_dtype": "float32",
|
| 33 |
+
"transformers_version": "4.51.3",
|
| 34 |
+
"type_vocab_size": 1,
|
| 35 |
+
"use_cache": true,
|
| 36 |
+
"vocab_size": 250002
|
| 37 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ce1483c47d31ce657985d4dfe229ef16a1326faa195dc0de24d2c52d60813ac5
|
| 3 |
+
size 2271071852
|
sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
|
| 3 |
+
size 5069051
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf44dabfaa82b1276a7af64a2ea2c76c047d560cf7bfb5711d6135382372c93d
|
| 3 |
+
size 17083153
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 512,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"sp_model_kwargs": {},
|
| 54 |
+
"tokenizer_class": "XLMRobertaTokenizer",
|
| 55 |
+
"unk_token": "<unk>"
|
| 56 |
+
}
|