Table-to-Text Generation on DART

12개 결과 · ⬇ CSV · JSON

METEOR

Rank	Model	METEOR	BLEU	BERT	BLEURT	Mover	TER	FactSpotter	Paper	Code	Year
1	T5B Baseline	0.4074	48.47	0.9505	0.6749	–	–	96.65	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
2	FactT5B	0.4072	48.37	0.9505	0.6743	–	–	97.60	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
3	JointGT Baseline	0.4043	47.51	0.9492	0.6733	–	–	95.86	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
4	FactJointGT	0.4032	47.39	0.9492	0.6726	–	–	97.25	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
5	HTLM (fine-tuning)	0.39	47.2	0.94	0.4	0.51	0.44	–	HTLM: Hyper-Text Pre-Training and Prompting of Language Models		2021
5	GPT-2-Large (fine-tuning)	0.39	47.0	0.94	0.4	0.51	0.46	–	HTLM: Hyper-Text Pre-Training and Prompting of Language Models		2021
7	T5B Baseline	0.4074	48.47	0.9505	0.6749	–	–	96.65	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
8	FactT5B	0.4072	48.37	0.9505	0.6743	–	–	97.60	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
9	JointGT Baseline	0.4043	47.51	0.9492	0.6733	–	–	95.86	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
10	FactJointGT	0.4032	47.39	0.9492	0.6726	–	–	97.25	FactSpotter: Evaluating the Factual Faithfulness of Graph-to-Text Generation	guihuzhang/FactSpotter	2023
11	HTLM (fine-tuning)	0.39	47.2	0.94	0.4	0.51	0.44	–	HTLM: Hyper-Text Pre-Training and Prompting of Language Models		2021
11	GPT-2-Large (fine-tuning)	0.39	47.0	0.94	0.4	0.51	0.46	–	HTLM: Hyper-Text Pre-Training and Prompting of Language Models		2021