OpenCodePapers

relation-classification-on-tacred-1

Relation Classification
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeF1ModelNameReleaseDate
DeepStruct: Pretraining of Language Models for Structure Prediction✓ Link36.1Deepstruct zero-shot2022-05-21
Zero-Shot Information Extraction as a Unified Text-to-Triple Translation✓ Link49.2DeepEx (zero-shot top-1)2021-09-23
Structured Prediction as Translation between Augmented Natural Languages✓ Link61.9TANL (multi-task)2021-01-14
ERNIE: Enhanced Language Representation with Informative Entities✓ Link66.0BERT2019-05-17
Graph Convolution over Pruned Dependency Trees Improves Relation Extraction✓ Link66.4C-GCN2018-09-26
ERNIE: Enhanced Language Representation with Informative Entities✓ Link68.0ERNIE2019-05-17
SpanBERT: Improving Pre-training by Representing and Predicting Spans✓ Link70.8SpanBERT2019-07-24
K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters✓ Link71.3RoBERTa2020-02-05
Matching the Blanks: Distributional Similarity for Relation Learning✓ Link71.5MTB Baldini Soares et al. (2019)2019-06-07
Knowledge Enhanced Contextual Word Representations✓ Link71.5KnowBERT2019-09-09
KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation✓ Link71.7KEPLER2019-11-13
Structured Prediction as Translation between Augmented Natural Languages✓ Link71.9TANL2021-01-14
K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters✓ Link72.0K-Adapter2020-02-05
LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention✓ Link72.7LUKE 483M2020-10-02
DeepStruct: Pretraining of Language Models for Structure Prediction✓ Link74.9DeepStruct multi-task2022-05-21
Zero-Shot Information Extraction as a Unified Text-to-Triple Translation✓ Link76.4DeepEx (zero-shot top-10)2021-09-23
DeepStruct: Pretraining of Language Models for Structure Prediction✓ Link76.8DeepStruct multi-task w/ finetune2022-05-21
Check out how to contribute new results.