MediChat is an advanced medical chatbot designed to assist with clinical queries and provide information based on medical literature. It leverages state-of-the-art models and embeddings to deliver accurate and reliable responses.

Provides information based on extensive training on medical literature.

Utilizes the Llama-2-7B-Chat model from Hugging Face.

Incorporates Chroma DB for efficient data retrieval.

Uses sentence-transformers/all-MiniLM-L6-v2 embeddings from Hugging Face.

• Source: Link
• Description: An open-source large language model optimized for chat-based interactions, capable of understanding and generating human-like text.

A vector database that allows for efficient storage and retrieval of embeddings.

• Source: Link
• Description: A small, fast, and high-quality embedding model that provides dense vector representations of text.

MediChat is trained on embeddings derived from the following medical books:

• Clinical Emergency Medicine (PDFDrive.com)
• Current Essentials of Medicine
• Gale Encyclopedia of Medicine Vol. 4 (N-S)

To set up MediChat, follow these steps:

Clone the Repository:

• git clone https://github.com/Gyanbardhan/MediChat.git
• cd MediChat

• pip install -r requirements.txt

• Llama-2-7B-Chat model: Download from Hugging Face and place it in the models directory.
• Sentence-Transformers embeddings: Download from Hugging Face and place them in the embeddings directory.
• Set Up Chroma DB

• python app.py