ChristopherLi/voyage-student-1-5b-dayfill
ChristopherLi/voyage-student-1-5b-dayfill is a 1.5 billion parameter language model distilled from Qwen2.5-1.5B-Instruct, specifically fine-tuned for travel planning tasks. It specializes in generating day-by-day itineraries and is optimized for local serving via Ollama. This model is designed to work as a student agent within a desktop travel planning application, focusing on structuring travel plans based on tool trajectories.
Loading preview...
Overview
ChristopherLi/voyage-student-1-5b-dayfill is a specialized 1.5 billion parameter language model designed for local travel itinerary drafting. It is a "student" model distilled from Qwen/Qwen2.5-1.5B-Instruct, specifically fine-tuned for use within the Local LLM AI Agent for Travel Planning desktop application.
Key Capabilities
- Travel Itinerary Generation: Optimized for creating detailed day-by-day travel plans.
- Tool Trajectory Integration: Fine-tuned using supervised fine-tuning (SFT) on Trip.Planner-style tool trajectories, including route proposal and
plan_tripactions. - Local Deployment: Packaged with Q4_K_M GGUF quantization, making it suitable for efficient local serving with Ollama and
llama.cpp. - Day-Fill Itinerary Structure: Specifically trained to excel at filling out daily schedules within a travel plan.
Good For
- Developers building local AI agents for travel planning that require structured itinerary outputs.
- Applications needing a compact, specialized model for generating day-by-day travel schedules.
- Use cases where the model interacts with external tools to gather travel information (e.g., Trip.com scrapes) and then structures that information into a coherent plan.
This model is intended for drafting itineraries and is not a booking engine; it relies on live tools for accurate pricing and URLs.