Relational Intelligence MedGemma (Merged)
A fine-tuned version of Google's MedGemma-4B-IT, trained on person-centred nursing language based on the Foundation of Nursing Studies (FONS) principles.
🏥 Model Description
This model was fine-tuned using Unsloth to enhance relational intelligence in healthcare conversations. It focuses on:
- Person-centred language - Compassionate, respectful communication
- Restorative supervision - Supporting nursing staff wellbeing
- Clinical reasoning - Grounded in medical knowledge from MedGemma
💻 Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained(
"NurseCitizenDeveloper/relational-intelligence-medgemma-merged",
torch_dtype="auto",
device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(
"NurseCitizenDeveloper/relational-intelligence-medgemma-merged"
)
messages = [{"role": "user", "content": "What is person-centred care?"}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0]))
📋 Training Details
- Base Model: google/medgemma-4b-it
- Fine-tuning Method: LoRA via Unsloth
- Training Data: FONS person-centred nursing language corpus
- Merged: 16-bit merged weights
⚠️ Disclaimer
This model is for educational and supportive purposes only. It does not provide clinical advice and should not replace professional medical judgment.
👨🏾⚕️ Author
Created by NurseCitizenDeveloper - Advancing AI in Nursing Education.
- Downloads last month
- 17