jastorj/snowflake_arctic_text2sql_r1_7b-nl2sqlpp-16bit-v5.7.5_phase_2-cw-32K

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:May 7, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

jastorj/snowflake_arctic_text2sql_r1_7b-nl2sqlpp-16bit-v5.7.5_phase_2-cw-32K is a 7.6 billion parameter model, fine-tuned from Snowflake/Arctic-Text2SQL-R1-7B, specifically designed for Text-to-SQL generation. It leverages the NL2SQL++ v8 dataset with code-with-thought reasoning and was fine-tuned using LoRA with Unsloth, featuring 16-bit merged weights. This model excels at converting natural language queries into valid SQL++ queries, including complex scenarios involving joins, aggregations, and schema-aware logic, making it ideal for database interaction and analytics.

Loading preview...

Snowflake/Arctic-Text2SQL-R1-7B Fine-tuned for NL2SQL++ v8

This model is a specialized 7.6 billion parameter language model, fine-tuned from the Snowflake/Arctic-Text2SQL-R1-7B base model. Its primary function is Text-to-SQL generation, specifically optimized for the SQL++ query language.

Key Capabilities

  • Accurate SQL++ Generation: Translates natural language queries into precise and valid SQL++ statements, adhering to specific schema constraints and SQL++ syntax rules.
  • Code-with-Thought Reasoning: Trained on the NL2SQL++ v8 dataset, which incorporates "code-with-thought" reasoning, enabling the model to generate more robust and logically sound queries by simulating a step-by-step thought process.
  • Schema-Aware Querying: Utilizes provided database schema to generate contextually relevant and executable queries, handling complex relationships, aggregations, and data types.
  • Efficient Fine-tuning: Leverages LoRA (Low-Rank Adaptation) with Unsloth for efficient fine-tuning, resulting in a performant model with 16-bit merged weights.

Good for

  • Automating Database Interactions: Ideal for applications requiring natural language interfaces to SQL++ databases, such as analytical tools, data exploration platforms, and business intelligence dashboards.
  • Complex Query Generation: Particularly effective for generating intricate SQL++ queries that involve multiple joins, conditional logic, aggregations (e.g., COUNT, ARRAY_AGG), and handling of missing/null values.
  • Developers and Data Analysts: Streamlines the process of writing SQL++ queries by allowing users to express their data needs in plain English, reducing manual query construction and potential errors.

Finetunes
2 models