You can deploy this model and start earning money today!

AliMurtaza-096/
qwen2.5-7b-medical-instruct

Name: qwen2.5-7b-medical-instruct
Availability: InStock

Views

No views yet

Qwen 2.5 7B Medical Front Desk Assistant

Fine-tuned version of Qwen2.5-7B-Instruct for medical front desk conversations, trained on 130 examples of clinic administrative tasks.

Model Details

Base Model: unsloth/Qwen2.5-7B-Instruct-bnb-4bit
Training Framework: Unsloth + LoRA
Training Steps: 80
Final Loss: 1.047
Training Data: 130 medical front desk conversations
Use Case: Medical clinic front desk assistant (scheduling, insurance, general clinic information)
LoRA Rank: 16
Target Modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj

Training Data

The model was fine-tuned on 130 conversations covering:

Appointment scheduling and rescheduling
Insurance verification
Office hours and location information
COVID-19 policies
General clinic information
Patient intake procedures

Usage

SafeTensors Format (Transformers)

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("AliMurtaza-096/qwen2.5-7b-medical-instruct")
tokenizer = AutoTokenizer.from_pretrained("AliMurtaza-096/qwen2.5-7b-medical-instruct")

messages = [
    {"role": "system", "content": "You are a friendly, professional front-desk assistant for Northview Medical Associates."},
    {"role": "user", "content": "I need to reschedule my appointment."}
]

inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
outputs = model.generate(inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0]))

GGUF Format (Ollama)

Download the qwen2.5-7b-medical-q4_k_m.gguf file (4.4GB) and create a Modelfile:

FROM ./qwen2.5-7b-medical-q4_k_m.gguf

TEMPLATE """<|im_start|>system
{{ .System }}<|im_end|>
<|im_start|>user
{{ .Prompt }}<|im_end|>
<|im_start|>assistant
"""

PARAMETER stop "<|im_start|>"
PARAMETER stop "<|im_end|>"

SYSTEM """You are a friendly, professional front-desk assistant for a medical clinic. You handle scheduling, insurance, and general clinic information. You NEVER diagnose or give medical treatment advice."""

Then:

bash

ollama create qwen2.5-medical -f Modelfile
ollama run qwen2.5-medical "What are your office hours?"

Example Responses

User: "Can I come in for a checkup on Saturday?"

Assistant: "Our office is open Monday through Friday from eight AM to five PM. We're closed on weekends and major holidays. Can I help you set up an appointment for next week?"

User: "I have an injection scheduled for tomorrow morning but I just tested positive for Covid. What should I do?"

Assistant: "I'm sorry to hear that. We need all patients to be negative on the day of their procedure. Please call us right away and we'll cancel your appointment and reschedule once you're well. In the meantime, please stay home and follow CDC guidelines."

Limitations

Does NOT provide medical advice or diagnoses - This model is trained exclusively for administrative front desk tasks
Trained specifically for front desk administrative tasks - May not generalize well to other medical contexts
Small training dataset - 130 examples may not cover all edge cases in real-world scenarios
Requires proper system prompt - Best results when used with the specified system prompt

Model Files

SafeTensors (15GB): Full precision model sharded across 4 files
- model-00001-of-00004.safetensors
- model-00002-of-00004.safetensors
- model-00003-of-00004.safetensors
- model-00004-of-00004.safetensors
GGUF Q4_K_M (4.4GB): Quantized model for efficient inference with Ollama/llama.cpp
- qwen2.5-7b-medical-q4_k_m.gguf

Training Details

Optimizer: AdamW 8-bit
Learning Rate: 2e-4
Scheduler: Linear
Batch Size: 2 per device
Gradient Accumulation: 4 steps
Effective Batch Size: 8
Warmup Steps: 5
Weight Decay: 0.01
Max Sequence Length: 4096
Precision: BF16 (on GPU with BF16 support)

Hardware

GPU: NVIDIA GeForce RTX 4090 (24GB VRAM)
Training Time: ~90 seconds for 80 steps

License

This model inherits the license from the base Qwen2.5-7B-Instruct model.

Acknowledgments

Fine-tuned using Unsloth for efficient training
Base model: Qwen/Qwen2.5-7B-Instruct
Quantized to GGUF using llama.cpp

Citation

If you use this model, please cite:

bibtex

@misc{qwen2.5-7b-medical-instruct,
  author = {Ali Murtaza},
  title = {Qwen 2.5 7B Medical Front Desk Assistant},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\\url{https://huggingface.co/AliMurtaza-096/qwen2.5-7b-medical-instruct}},
}

⚠️ Medical Disclaimer: This AI model is designed for administrative front desk tasks only and should NOT be used for medical diagnosis, treatment recommendations, or clinical decision-making. Always consult qualified healthcare professionals for medical advice.