LLM Reference

NSQL Llama 2 7B

About

The NSQL Llama 2 7B model is an open-source, autoregressive large language model tailored for SQL query generation. It is based on Meta's Llama 2 7B and undergoes pre-training on a vast SQL dataset and fine-tuning with text-to-SQL pairs, enhancing its ability to convert natural language prompts into precise SQL queries. The model excels in long-context scenarios and complex queries, including multi-table joins and nested queries. It’s readily accessible on Hugging Face, encouraging community participation. Nonetheless, challenges remain in date-related calculations and might require further fine-tuning for specific domains.

Capabilities

MultimodalFunction CallingTool UseJSON Mode

Specifications

FamilyNSQL
Parameters7B
ArchitectureDecoder Only
Specializationgeneral