Natural Language (NL) to SQL 1,000-Pair AI Training Dataset - Food (CSV, JSON)

Natural language (NL) to SQL:1000 verified training pairs dataset for Natural Language to SQL translation. It can be used for a food - discovery search or question-answering systems, agents and (gen AI apps).

1000 new pairs

• Download CSV

• Download TXT

• Download JSON.

This training dataset is meant for building intelligent search or question-answering systems using Large Language Models (LLMs) which rely on training data for prompt engineering or fine-tuning.

 

A dataset of this magnitude might be used internally for training AI models. We constructed (knowledge entities both by generating synthetic pairs and by crowdsourcing questions/answers for a general food-knowledge graph. 

While the datasets of this scale might exist in the food domain, but may not be specifically tied to the exact purpose.

The dataset can be used for/in rapid development as a light-weight training for agents/ gen AI apps as well as can be added to MCPs and RAGs,