XGenerationLab/XiYanSQL-QwenCoder-14B-2502
The XiYanSQL-QwenCoder-14B-2502 model by XGenerationLab is a 14.8 billion parameter language model specifically designed for text-to-SQL tasks, supporting a 32768 token context length. It excels at generating SQL queries from natural language across multiple dialects including SQLite, PostgreSQL, and MySQL. This model demonstrates strong performance on benchmarks like BIRD and Spider, making it suitable for direct use or as a fine-tuning base for SQL generation applications.
Loading preview...
XiYanSQL-QwenCoder-14B-2502: Specialized Text-to-SQL Model
The XiYanSQL-QwenCoder series, developed by XGenerationLab, focuses on advancing large language models for text-to-SQL generation. This 14.8 billion parameter variant is part of a family that includes 3B, 7B, and 32B models, all optimized for converting natural language into SQL queries.
Key Capabilities
- High Performance: The XiYanSQL-QwenCoder-32B model achieved a 69.03% EX score on the BIRD TEST set, setting a new state-of-the-art for single fine-tuned models, with the 14B model also showing strong competitive results.
- Multi-Dialect Support: It natively supports major SQL dialects such as SQLite, PostgreSQL, and MySQL, making it versatile for various database environments.
- Flexible Usage: The model can be directly applied to text-to-SQL tasks or serve as an effective base for further fine-tuning to specific SQL generation needs.
- Benchmark Excellence: Evaluated on BIRD and Spider benchmarks, the models demonstrate robust SQL generation capabilities across different schema formats (M-Schema and original DDL).
When to Use This Model
- Text-to-SQL Applications: Ideal for developers building applications that require converting natural language questions into executable SQL queries.
- Database Interaction: Useful for automating database interactions and simplifying data retrieval for non-technical users.
- Fine-tuning Base: Provides a strong foundation for researchers and developers looking to fine-tune a specialized SQL generation model for unique datasets or domain-specific requirements.