rishiraj/smolified-tiny-text-to-sql

TEXT GENERATIONConcurrent Unit Cost:1Model Size:0.3BQuant:BF16Context Size:32kPublished:Feb 14, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

The rishiraj/smolified-tiny-text-to-sql is a 0.3 billion parameter Domain Specific Language Model (DSLM) from the Smolify Foundry, synthetically distilled from larger reasoning engines. Optimized for text-to-SQL translation, this micro-architecture is designed for high-efficiency deployment on edge hardware or in low-VRAM environments. It specializes in converting natural language questions into valid SQLite queries, making it suitable for embedded database applications.

Loading preview...

Overview

The rishiraj/smolified-tiny-text-to-sql is a 0.3 billion parameter Domain Specific Language Model (DSLM) developed by rishiraj through the Smolify Foundry. This model is synthetically distilled from state-of-the-art reasoning engines, resulting in a highly efficient architecture. It is specifically optimized for deployment on edge hardware (CPU/NPU) or in environments with limited VRAM.

Key Capabilities

  • Text-to-SQL Translation: Specializes in converting natural language queries into valid SQLite statements.
  • High Efficiency: Designed for low-resource environments due to its small parameter count (270M parameter class).
  • Optimized Deployment: Features 4-bit quantized and FP16 mixed optimization for efficient inference.
  • Domain Specific: Focused on SQL generation, making it highly effective for database interaction tasks.

Good For

  • Applications requiring efficient text-to-SQL capabilities on resource-constrained devices.
  • Integrating natural language database querying into embedded systems.
  • Use cases where a compact, specialized model for SQL generation is preferred over larger, general-purpose LLMs.