CareMinds-AI

🧠 CareMinds-AI — Medical AI Assistant

CareMinds-AI is a lightweight, offline-capable medical AI system built fine-tuned with domain-specific healthcare data. It combines:

• ✅ LLM (TinyLLaMA + LoRA fine-tuning)
• ✅ Structured Data Analytics (CSV + Pandas)
• ✅ RAG-style context retrieval (lightweight)
• ✅ Patient record lookup system

---

📌 Model Details

📖 Model Description

CareMinds-AI is a hybrid AI system designed to:

• Answer general medical questions
• Analyze structured healthcare datasets
• Retrieve patient-specific records
• Perform real-time analytics using natural language

It operates fully offline, without requiring external APIs.

---

👨‍💻 Developed by

Vedhamani Prabakar A

🧠 Model Type

• Base Model: CareMinds-AI (1.1B parameters)
• Fine-tuning: LoRA (PEFT)
• Architecture: Transformer-based causal language model

🌐 Language(s)

• English

---

🔗 Model Sources

• GitHub Repository:
  https://github.com/VedhamaniprabakarA/CareMinds-AI.git

---

🚀 Uses

✅ Direct Use

CareMinds-AI can be used as:

• 🧠 Medical chatbot (offline)
• 📊 Healthcare data analyzer
• 🏥 Patient record retrieval system
• 🔍 Natural language query engine

---

🔄 Downstream Use

• Hospital management systems
• Healthcare dashboards
• Clinical data assistants
• AI-powered analytics tools

---

🧠 Recommendations

• Use alongside verified medical systems
• Add RAG with trusted medical sources for production
• Validate outputs before real-world use

---

⚙️ How to Use the Model

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_path = "./CareMinds-AI"

model = AutoModelForCausalLM.from_pretrained(model_path)
tokenizer = AutoTokenizer.from_pretrained(model_path)

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

prompt = """### Instruction:
Explain about diabetes

### Response:
"""
inputs = tokenizer(prompt, return_tensors="pt").to(device)

outputs = model.generate(**inputs, max_new_tokens=150)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))


###test the model:
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("vedhamani/CareMinds-AI")
tokenizer = AutoTokenizer.from_pretrained("vedhamani/CareMinds-AI")

prompt = "### Instruction:\nExplain about diabetes\n\n### Response:\n"

inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)

print(tokenizer.decode(outputs[0]))