Explore high-quality datasets for your AI and machine learning projects.
The gretelai/synthetic_text_to_sql dataset is a high‑quality synthetic Text‑to‑SQL sample dataset generated using Gretel Navigator, containing 105,851 records, split into 100,000 training records and 5,851 test records. The dataset covers 100 different domains and includes various SQL tasks such as data definition, retrieval, manipulation, analysis, and reporting. Additionally, the dataset provides natural‑language explanations of SQL queries and contextual tags to optimize model training. The dataset quality is evaluated using LLM‑as‑a‑judge techniques, showing excellent performance in SQL standard compliance, correctness, and instruction adherence.
