Hi Everyone,

This is my new project based on Gen AI, named as Thought Therapist.

It is a streamlit website, which has three sections:

1. Thought Therapist
2. FAQ (Frequently Asked Questions)
3. Know more about developer.

Thought Therapist

It is a chatbot, which chats with a person facing mental health issues. It is a fine-tuned model on LAMINI, which is then customly integrated with Langchain and streamlit to create the working web page. It is based on LLM Meta-Llama-3.1-8B. It has the feature of keeping the context of chat history in order to provide valuable feedback.

FAQ

It is a question-answering prompt, which answer frequenty asked questions based on mental health based on a CSV file containing the FAQ's. It used Vector Data Base to store the vector embeddings and it used Retrieval-Augmented Generation(RAG) in order to produce output from the given prompt and the context. It used EDEN AI for embeddings and GROQ AI as LLM provider.

Know more about developer

It is a question-answering prompt, which answer any queries about the developer. Here, I have used Graph Data Base by using neo4j. I have made a graph Data base about me and then I have integrated this database with Langchain to answer the queries based on the information in database. Here, you can ask questions like

1. "What is your name? "

2. "Tell me about your experience?"

3. "Which Languages do you know?"

4. "Give me a introduction of yourself."

The website is hosted on streamlit cloud service as https://thoughttherapist.streamlit.app/

Files and Folders details

1. .streamlit: It is a folder that contains pages information on streamlit i.e. details and structure of three pages.

2. FAQ.CSV: It contains the frequently asked questions in CSV format.

3. FAQ_Vector_DB.py : It contains a python program, which was used to create the vector embeddings for the FAQ.csv and store them in a database i.e. FAQ_db.

4. FAQ_db : Database for embeddings.

5. Model_Fine-tuning: It contains a python program which fine tunes the model on LAMINI based on data provided.

6. Therapist.py: The first section of the website.

7. FAQ.py: The second section of the website.

8. about.py: The third section of the website.

9. app.py L The website starts from this python file.

10. GraphDB_Mukul: Contains the structure of database created for section Know more about developer. (Just for reference, this image isn't used anywhere.)

Please find the demo at: https://drive.google.com/file/d/1H8CtNSd8t5JhvqaDO73i4LuchCwAWKEG/view?usp=drive_link