roberta-large-semeval2012-v2-mask-prompt-e-nce Open-source Model - Help You Solve Lexical Relation Classification and Analogy Problems

Roberta Large Semeval2012 V2 Mask Prompt E Nce

Developed by research-backup

A relation understanding model fine-tuned based on RoBERTa-large, specializing in lexical relation classification and analogy question answering

Text Embedding

Transformers

#Relation Embedding Extraction #Lexical Relation Classification #Analogy Question Answering

Downloads 16

Release Time : 8/18/2022

Model Overview

This model is based on the RoBERTa-large architecture and fine-tuned on the SemEval2012 relation similarity dataset. It is primarily used for relation understanding tasks, including lexical relation classification and analogy question answering.

Model Features

Efficient Relation Understanding

Performs excellently on multiple relation understanding tasks, especially achieving high F1 scores in lexical relation classification tasks

Multi-Task Adaptability

Capable of handling various relation understanding tasks, including relation mapping, lexical relation classification, and analogy question answering

Fine-Tuned Based on RoBERTa-large

Utilizes the powerful RoBERTa-large pre-trained model as a foundation, achieving outstanding performance through task-specific fine-tuning

Model Capabilities

Lexical relation classification

Analogy question answering

Relation mapping

Relation similarity calculation

Use Cases

Education

Analogy Question Answering

Used to answer analogy questions similar to those in SAT exams

Achieves an accuracy of 60.96% on the full SAT version

Natural Language Processing

Lexical Relation Classification

Classifies semantic relations between words

Achieves an F1 score of 92.65% on the BLESS dataset

🚀 relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce

RelBERT is fine-tuned from roberta-large on relbert/semeval2012_relational_similarity_v2. The fine-tuning process is carried out via the RelBERT library. For more details, please refer to the repository. This model achieves excellent results in relation understanding tasks.

🚀 Quick Start

RelBERT fine-tuned from roberta-large on relbert/semeval2012_relational_similarity_v2. It can be used to perform various relation understanding tasks.

✨ Features

This model can be used for multiple relation understanding tasks, including analogy questions, lexical relation classification, and relation mapping. It has been fine-tuned on specific datasets to achieve high performance.

📦 Installation

You can install the necessary library via pip:

pip install relbert

💻 Usage Examples

Basic Usage

from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce")
vector = model.get_embedding(['Tokyo', 'Japan'])  # shape of (1024, )

📚 Documentation

Results on Relation Understanding Tasks

Task	Dataset	Metric	Value
Analogy Question	relbert/analogy_questions	Accuracy on SAT (full)	0.6096256684491979
Analogy Question	relbert/analogy_questions	Accuracy on SAT	0.6112759643916914
Analogy Question	relbert/analogy_questions	Accuracy on BATS	0.7576431350750417
Analogy Question	relbert/analogy_questions	Accuracy on U2	0.5964912280701754
Analogy Question	relbert/analogy_questions	Accuracy on U4	0.6087962962962963
Analogy Question	relbert/analogy_questions	Accuracy on Google	0.878
Lexical Relation Classification	relbert/lexical_relation_classification	Micro F1 score on BLESS	0.9264728039777008
Lexical Relation Classification	relbert/lexical_relation_classification	Micro F1 score on CogALexV	0.8720657276995305
Lexical Relation Classification	relbert/lexical_relation_classification	Micro F1 score on EVALution	0.7074756229685807
Lexical Relation Classification	relbert/lexical_relation_classification	Micro F1 score on K&H+N	0.9625095638867636
Lexical Relation Classification	relbert/lexical_relation_classification	Micro F1 score on ROOT09	0.9022250078345346
Relation Mapping	relbert/relation_mapping	Accuracy on Relation Mapping	0.8457142857142858

Training Hyperparameters

The following hyperparameters were used during training:

Parameter	Value
model	roberta-large
max_length	64
mode	mask
data	relbert/semeval2012_relational_similarity_v2
template_mode	manual
template	I wasn’t aware of this relationship, but I just read in the encyclopedia that is ’s
loss_function	nce_logout
temperature_nce_constant	0.05
temperature_nce_rank	{'min': 0.01, 'max': 0.05, 'type': 'linear'}
epoch	29
batch	128
lr	5e-06
lr_decay	False
lr_warmup	1
weight_decay	0
random_seed	0
exclude_relation	None
n_sample	640
gradient_accumulation	8

The full configuration can be found at fine-tuning parameter file.

🔧 Technical Details

The model is fine-tuned from roberta-large using the RelBERT library. The fine-tuning process involves optimizing the model on the relbert/semeval2012_relational_similarity_v2 dataset with specific hyperparameters.

📄 License

No license information provided in the original document.

📖 Reference

If you use any resource from RelBERT, please consider to cite our paper.

@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
    title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
    author = "Ushio, Asahi  and
      Schockaert, Steven  and
      Camacho-Collados, Jose",
    booktitle = "EMNLP 2021",
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
}

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご