jarminraws/hotel-llm-search

TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 22, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

The jarminraws/hotel-llm-search model is a 2 billion parameter Qwen3-1.7B-based language model fine-tuned for extracting structured hotel-search parameters from natural language queries. It specializes in converting user input into a strict JSON format, including details like destination, dates, guest counts, and human-readable filter phrases. With a 32768-token context length, this model is optimized for precise entity extraction in hotel booking applications.

Loading preview...

Overview

The jarminraws/hotel-llm-search model is a 2 billion parameter language model built upon the Qwen3-1.7B architecture, specifically fine-tuned for hotel entity extraction. Its primary function is to parse natural language hotel search queries and output a structured JSON object containing relevant search parameters.

Key Capabilities

  • Structured JSON Output: Converts free-form text queries into a predefined JSON schema, including fields like destination, locality, hotelName, checkin/checkoutDate, adult/room/child/infantCount, sortCriteria, min/maxPrice, filters, deepSearch, isNearMe, and resetAction.
  • Human-Readable Filters: Unlike many models that output codes, this model emits filters as human-readable phrases (e.g., "swimming pool, pet friendly"), allowing for downstream resolution to specific production codes.
  • Date Math Integration: Designed to process queries that include relative date references, with an explicit today date (DDMMYYYY) input for accurate date calculation.
  • High Context Length: Features a 32768-token context window, enabling it to handle complex and detailed search queries.

Good For

  • Building Hotel Search Interfaces: Ideal for powering conversational AI or search bars in hotel booking platforms that require precise extraction of user intent.
  • Reducing Downstream Complexity: By providing human-readable filter phrases, it simplifies the integration with existing backend systems that map phrases to internal codes.
  • Ensuring Data Consistency: Adheres to a strict prompt/schema contract (contract.py), ensuring consistent output format crucial for reliable system integration.

Finetunes
1 models