WikiSQL: Natural Language Questions and SQL Queries for Model Fine-Tuning

A dataset processed from salesforce/WikiSQL, used to demonstrate the fine-tuning function of LLMs in the mlx_lm library. It was created by mlx-community and last updated on September 20, 2024. The dataset likely contains pairs of natural language questions and corresponding SQL queries.

Use Cases

• Fine-tuning language models for text-to-SQL conversion based on the described question-query pairs.
• Benchmarking model performance on natural language to structured query language tasks.
• Training models for question answering over relational databases using the SQL query structure.

Strengths

• Processed from the established salesforce/WikiSQL dataset, suggesting a known foundation.
• Specifically designed for demonstrating LLM fine-tuning in the mlx_lm library.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Processed from salesforce/WikiSQL.

Collection Method

Processing method is not specified in the provided description.

Freshness

Last updated 2024-09-20 06:52:13; freshness should be verified.