Mayank-AI: Medical AI Assistant Model

📋 Model Overview

Mayank-AI is a specialized artificial intelligence model designed for Indian pharmaceutical and medical applications, trained on comprehensive Indian medicines datasets. This model leverages supervised learning techniques built on GPT-2 transformer architecture to provide accurate and contextually relevant information about Indian medicines, their compounds, uses, and related medical information.

🔍 Model Details

Model Description

Developed by: Mayank Malviya
Model Type: GPT-2 based Transformer for Indian Medical/Pharmaceutical Applications
Language(s): English (with Indian medical terminology and drug names)
License: Apache-2.0
Domain: Indian Pharmaceuticals & Medicine Information
Primary Use: Indian medicine information, drug compound analysis, symptom mapping, prescription guidance

Key Features

✅ Indian medicines database knowledge
✅ Drug compound information and analysis
✅ Symptom-to-medicine mapping
✅ Prescription guidance and recommendations
✅ Disease diagnosis assistance
✅ Indian pharmaceutical market insights
✅ Medicine availability and alternatives

🚀 Quick Start

Installation

pip install transformers torch

Basic Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Load model and tokenizer
model_name = "Mayank-22/Mayank-AI"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Example queries for Indian medicines
query1 = "What is the composition of Crocin tablet?"
query2 = "Which medicine is used for fever and headache?"
query3 = "What are the side effects of Paracetamol?"
query4 = "Medicines available for diabetes in India"

# Process query
inputs = tokenizer.encode(query1, return_tensors="pt")

# Generate response
with torch.no_grad():
    outputs = model.generate(
        inputs,
        max_length=512,
        num_return_sequences=1,
        temperature=0.7,
        do_sample=True,
        pad_token_id=tokenizer.eos_token_id
    )

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

Advanced Usage

# For more controlled generation about Indian medicines
def generate_medicine_response(question, max_length=256):
    prompt = f"Indian Medicine Query: {question}\nResponse:"
    inputs = tokenizer.encode(prompt, return_tensors="pt")
    
    outputs = model.generate(
        inputs,
        max_length=max_length,
        num_return_sequences=1,
        temperature=0.6,
        do_sample=True,
        top_p=0.9,
        repetition_penalty=1.1
    )
    
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response.split("Response:")[-1].strip()

# Example usage
question = "What are the uses of Azithromycin tablets available in India?"
answer = generate_medicine_response(question)
print(answer)

📊 Performance & Capabilities

Supported Medical Areas

Indian Pharmaceuticals: Comprehensive database of medicines available in India
Drug Compounds: Active ingredients, chemical compositions, formulations
Symptom Analysis: Symptom-to-medicine mapping and recommendations
Disease Information: Common diseases and their standard treatments in India
Prescription Guidance: Dosage, administration, and usage instructions
Drug Interactions: Side effects and contraindications
Medicine Alternatives: Generic and branded medicine alternatives

Performance Metrics

Training Data: Indian medicines dataset with comprehensive drug information
Specialization: Focused on Indian pharmaceutical market and medicine availability
Coverage: Extensive database of Indian medicines, their compounds, and uses
Accuracy: High precision in Indian medicine information and drug compound details

⚠️ Important Medical Disclaimer

CRITICAL NOTICE: This model is for informational and educational purposes only. It should NOT be used as a substitute for professional medical advice, diagnosis, or treatment. Always consult with qualified healthcare providers for medical concerns.

Limitations & Risks

Not a replacement for medical professionals
May contain inaccuracies or outdated information
Should not be used for emergency medical situations
Requires human oversight for clinical applications
May have biases present in training data

🎯 Intended Use Cases

✅ Appropriate Uses

Indian pharmaceutical research and education
Medicine information lookup and comparison
Drug compound analysis and research
Symptom-to-medicine mapping assistance
Prescription guidance and dosage information
Medicine availability and alternatives research
Healthcare app development and integration

❌ Inappropriate Uses

Direct patient diagnosis
Emergency medical decisions
Prescription or treatment recommendations without medical supervision
Replacement for clinical judgment
Use without proper medical context

🔧 Technical Specifications

Model Architecture

Base Architecture: GPT-2 Transformer model
Fine-tuning: Supervised learning on Indian medicines dataset
Context Length: Standard GPT-2 context window
Training Approach: Domain-specific fine-tuning on pharmaceutical data

Training Details

Training Data: Indian medicines dataset including:
- Medicine names and brand information
- Drug compounds and chemical compositions
- Symptom-medicine mappings
- Prescription guidelines and dosages
- Disease-treatment associations
- Side effects and contraindications
Training Regime: Supervised fine-tuning on GPT-2 with pharmaceutical domain adaptation
Optimization: Adam optimizer with learning rate scheduling
Data Focus: Indian pharmaceutical market and medicine availability

📚 Datasets & Training

Training Data Sources

Comprehensive Indian medicines database
Drug compound and chemical composition data
Symptom-medicine relationship mappings
Prescription guidelines and dosage information
Disease-treatment associations
Medicine availability and market data

Data Preprocessing

Medicine name normalization and standardization
Drug compound data structure optimization
Symptom-medicine relationship mapping
Quality filtering and validation of pharmaceutical data
Indian market-specific data curation

🧪 Evaluation & Validation

Evaluation Metrics

Medicine Information Accuracy: Correctness of drug compound and usage information
Symptom Mapping Precision: Accuracy of symptom-to-medicine recommendations
Indian Market Relevance: Appropriateness for Indian pharmaceutical context
Safety Assessment: Risk evaluation for medicine information provision

Benchmark Performance

Indian Medicine Database: Comprehensive coverage of medicines available in India
Drug Compound Accuracy: High precision in chemical composition information
Symptom-Medicine Mapping: Effective symptom-to-treatment recommendations

🔄 Updates & Maintenance

This model is maintained and updated with:

Latest Indian medicine information
New drug approvals and market entries
Updated compound and formulation data
Enhanced symptom-medicine mappings

📖 Citation

If you use this model in your research, please cite:

@misc{mayank2024indianmedicines,
  title={Mayank-AI: Indian Medicines Information Model},
  author={Malviya, Mayank},
  year={2024},
  url={https://huggingface.co/Mayank-22/Mayank-AI},
  note={GPT-2 based model for Indian pharmaceutical information}
}

🤝 Contributing

Contributions to improve the model are welcome! Please:

Report issues with medicine information accuracy
Suggest new Indian medicines to include
Share feedback on drug compound data
Contribute to symptom-medicine mapping improvements

📞 Contact & Support

Model Author: Mayank Malviya
Repository: Mayank-22/Mayank-AI
Issues: Please report issues through the Hugging Face repository

📄 License

This model's license is not currently specified. Please check the repository or contact the author for licensing information.

🙏 Acknowledgments

Special thanks to the Indian pharmaceutical community, healthcare professionals, and medical researchers who contributed to the development and validation of this specialized model for Indian medicines.

Remember: This AI model is a tool to assist, not replace, medical professionals. Always prioritize patient safety and seek professional medical advice for healthcare decisions.

Mayank-22
/

Mayank-AI