This repository hosts a medical chatbot project developed using the Llama-2-7b-chat-hf model by Meta and implemented using the Langchain framework. The chatbot utilizes state-of-the-art LLM model and natural language processing (NLP) techniques to provide personalized healthcare support, including symptom analysis, treatment recommendations, and general health guidance.

• Advanced NLP Capabilities: The chatbot leverages cutting-edge technologies like Foundation models(LLMs) and NLP techniques for accurate understanding and response generation.
• Tailored Medical Assistance: Users can receive personalized assistance, including symptom analysis and treatment recommendations.
• Extensive Knowledge Base: The chatbot is equipped with a comprehensive knowledge base to provide reliable healthcare information.
• Improved Healthcare Access: The project aims to enhance accessibility to healthcare information, promoting proactive health management.

• Python 3.8 or above
• HuggingFace
• Langchain
• Gradio
• Dependencies specified in requirements.txt

1. Clone the repository:

git clone https://github.com/shishir-dwi/HealPal-Chatmate

2. Install dependencies:

pip install -r requirements.txt

3. Llama2 models require gated access to the model from hugging face.

Go to the hugging face > sign up > find model > fill form by Meta and then HF for the access > create a access token fro the model and store it in a file named "credentials.env"

4. Run the application using Gradio

1. Start the chatbot application by running HealPal Prototype-I.ipynbfile.
2. Interact with the chatbot by entering queries related to healthcare through Gradio interface.
3. Receive personalized assistance, symptom analysis, treatment recommendations, and general health guidance.

Contributions to the project are welcome. Please feel free to submit pull requests or open issues for any suggestions, bug reports, or feature requests.

This project is licensed under the MIT License.

• Meta for providing the Llama-2 model.

For any inquiries or support, please contact [email protected].