Text To Sql On Spider

评估指标

Exact Match Accuracy (Dev)

Exact Match Accuracy (Test)

Execution Accuracy (Test)

评测结果

各个模型在此基准测试上的表现结果

模型名称	Exact Match Accuracy (Dev)	Exact Match Accuracy (Test)	Execution Accuracy (Test)	Paper Title	Repository
T5-3B + PICARD	75.5	71.9	75.1	PICARD: Parsing Incrementally for Constrained Auto-Regressive Decoding from Language Models	-
MSc-SQL	-	-	84.7	MSc-SQL: Multi-Sample Critiquing Small Language Models For Text-To-SQL Translation	-
T5-3B+NatSQL+Token Preprocessing	69.4	-	78	Improving Generalization in Language Model-Based Text-to-SQL Semantic Parsing: Two Simple Semantic Boundary-Based Techniques	-
datagpt-sql-7B + InvalidSQL-Feedback	81.6	-	-	DataGpt-SQL-7B: An Open-Source Language Model for Text-to-SQL	-
SADGA + GAP	73.1	70.1	-	SADGA: Structure-Aware Dual Graph Aggregation Network for Text-to-SQL	-
MAPO + TABERTLarge (K = 3)	64.5	-	-	TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data	-
T5-SR	77.2	72.4	75.2	T5-SR: A Unified Seq-to-Seq Decoding Strategy for Semantic Parsing	-
DAIL-SQL + GPT-4 + Self-Consistency	74.4	-	86.6	Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation	-
code-davinci-002 175B (LEVER)	-	-	-	LEVER: Learning to Verify Language-to-Code Generation with Execution	-
PET-SQL	-	66.6	87.6	PET-SQL: A Prompt-Enhanced Two-Round Refinement of Text-to-SQL with Cross-consistency	-
RESDSQL-3B + NatSQL	80.5	72.0	79.9	RESDSQL: Decoupling Schema Linking and Skeleton Parsing for Text-to-SQL	-
MARLO + Claude 2.1	-	-	84.0	Learning Metadata-Agnostic Representations for Text-to-SQL In-Context Example Selection	-
DIN-SQL + GPT-4	-	60	85.3	DIN-SQL: Decomposed In-Context Learning of Text-to-SQL with Self-Correction	-
RYANSQL (BERT)	66.6	58.2	-	RYANSQL: Recursively Applying Sketch-based Slot Fillings for Complex Text-to-SQL in Cross-Domain Databases	-
Graphix-3B+PICARD	77.1	-	77.6	Graphix-T5: Mixing Pre-Trained Transformers with Graph-Aware Layers for Text-to-SQL Parsing	-
datagpt-sql-7B	80.3	-	-	DataGpt-SQL-7B: An Open-Source Language Model for Text-to-SQL	-
C3 + ChatGPT + Zero-Shot	-	-	82.3	C3: Zero-shot Text-to-SQL with ChatGPT	-
RATSQL + GAP	71.8	-	-	Learning Contextual Representations for Semantic Parsing with Generation-Augmented Pre-Training	-
XiYan-SQL	-	-	89.65	A Preview of XiYan-SQL: A Multi-Generator Ensemble Framework for Text-to-SQL	-
DELLM + GPT-4	-	-	-	Knowledge-to-SQL: Enhancing SQL Generation with Data Expert LLM	-

0 of 20 row(s) selected.