Built with: simula — schema-driven synthetic data generation with auditable taxonomy lineage.
An English synthetic dataset for training and evaluating systems that translate natural-language
shopping requests into narrow, atomic, database-queryable JSON. It contains 10,985 accepted
examples from a 13,000-attempt generation run. No accepted rows were trimmed from this release.
Each example pairs a realistic typed or spoken shopper query… See the full description on the dataset page:
https://huggingface.co/datasets/Ionio-ai/ecommerce-search-extraction.
