This dataset was generated using Distilabel and contains natural language questions paired with SQL queries. The model used is llama3.2:3b-instruct-fp16.
The generation environment was a test to use Ollama in combination with a VSCode devcontainer environment using uv to better control dependencies/reproducibility.
Inputs: Natural language questions
Outputs: Corresponding SQL queries
Generated via: johnmccabe/ollama_sql_review_test… See the full description on the dataset page:
https://huggingface.co/datasets/johnmccabe/ollama_sql_review_test.
