RadEstate-Qwen2.5-0.5B-IntakeAdapter

RadEstate Qwen2.5-0.5B Intake Adapter

This model is a LoRA fine-tuned version of Qwen2.5-0.5B-Instruct optimized for RadEstate intake question ingestion.

The model has been trained to better understand real-estate-related intake data and assist with extracting structured information from user inputs.

Model Details

• Base model: Qwen/Qwen2.5-0.5B-Instruct
• Fine-tuning method: LoRA (Low-Rank Adaptation)
• Quantization: 4-bit quantized model
• Training dataset size: ~2,000 examples
• Task: Real-estate intake question ingestion

Intended Use

This model is intended for:

• Real-estate intake processing
• Understanding domain-specific user responses
• Extracting structured information from natural language inputs
• Supporting RadEstate workflows

Training Details

The model was fine-tuned using LoRA on top of Qwen2.5-0.5B-Instruct.

Training approach:

• Parameter-efficient fine-tuning using LoRA
• Domain-specific instruction dataset
• Approximately 2,000 training examples

After fine-tuning, the model was quantized to 4-bit precision for reduced memory usage and efficient deployment.

Usage

This repository contains the fine-tuned model weights.

Example:

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "radcrew/RadEstate-Qwen2.5-0.5B-Intake"

tokenizer = AutoTokenizer.from_pretrained(model_id)

model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto"
)