hirmnm/qwen2.5-1.5B_rewriter
The hirmnm/qwen2.5-1.5B_rewriter is a 1.5 billion parameter language model, fine-tuned from Qwen/Qwen2.5-1.5B-Instruct, specifically designed for query rewriting. It specializes in converting raw user queries into structured JSON output, including canonical queries and rewrite types. This model is optimized for router query normalization and canonicalization tasks, offering a context length of 32768 tokens.
Loading preview...
Model Overview
The hirmnm/qwen2.5-1.5B_rewriter is a specialized 1.5 billion parameter language model, fine-tuned from the Qwen/Qwen2.5-1.5B-Instruct architecture. Its primary function is to act as a query rewriter, transforming unstructured user inputs into a standardized, machine-readable JSON format.
Key Capabilities
- Query Rewriting: Converts raw user queries into a strict JSON structure.
- Structured Output: Generates JSON with specific fields:
canonical_query,rewrite_type,is_multi_candidate, andsub_queries. - Fine-tuned Performance: Utilizes QLoRA for efficient fine-tuning, focusing on completion-only loss.
- Context Length: Supports a substantial context window of 32768 tokens.
Intended Use Cases
This model is particularly well-suited for applications requiring router query normalization and canonicalization. It can be integrated into systems where user queries need to be consistently formatted and categorized before further processing, such as in search engines, conversational AI, or data routing mechanisms. Its ability to output strict JSON ensures seamless integration into automated workflows.