johnnycpu/qwen2.5-3b-welfare-intake-lora

微光導航 LightGuide — Task A — IntakeAgent Slot Extraction

Qwen2.5-3B-Instruct fine-tuned with QLoRA (r=16, alpha=64) on 3300 synthetic Traditional Chinese welfare intake dialogues. Extracts 6 structured slots (age, city, household_members, monthly_household_income, disability_level, housing_status) from conversational input including Hokkien/Taiwanese mixed speech. Part of the 微光導航 (LightGuide) project for elderly welfare navigation.

Model Details

• Base Model: Qwen/Qwen2.5-3B-Instruct
• Fine-tuning Task: Task A — IntakeAgent Slot Extraction
• Project: 微光導航 (LightGuide) — AI-powered elderly welfare navigation system
• Competition: 大專生資訊服務創新競賽 (Taiwan University Information Service Innovation Competition)
• License: Apache 2.0

Usage

from transformers import AutoTokenizer, AutoModel
from peft import PeftModel  # if LoRA adapter

# Load from Hub
model = AutoModel.from_pretrained("johnnycpu/qwen2.5-3b-welfare-intake-lora")

Training Details

• Dataset: Custom Traditional Chinese welfare dialogue dataset
• Training Hardware: NVIDIA A100 (Google Colab)
• Framework: HuggingFace Transformers + PEFT

Citation

If you use this model, please cite the 微光導航 project.