llama + SQL + PyGWalker + Flask
A personally hosted chatbot that can answer questions based on the data it has. You will be able to add, view and delete files(PDFs, CSVs and TXTs) from the UI. NOTE: When a file is deleted the corresponding embeddings are also deleted, so be sure whether you don't want that data to be considered by the model. A maximum of 5 chat sessions can be saved at a time. You will be able to save, delete and export the chat when needed. Be sure to save the chat before switching to a new session. You can also toggle the background music on/off.
Upload the document, wait for the vectorization to complete, create a session, type your query and hit enter.
The application uses gpt4all-groovy model and can run on CPU (time taken to query the output varies depending on the hardware though).
You can download the model from here
# Clone the repository
git clone https://github.com/raghujhts13/private-chatbot.git
# change directory to where files are present
cd private-chatbot
# install the required libraries using the requirements.txt
pip install -r requirements.txt
# start the application
python app.pyconfigure the .env file MODEL_PATH - path where you have the model downloaded MODEL_TYPE - llama.cpp or gpt4all ( change the model accordingly) EMBEDDINGS_MODEL_NAME - embedding to use json_file_path - where the chat histories are stored OUTPUT_PATH - where a particular chat session can be stored
You can access the application at http://127.0.0.1:5000 (if this is the port where the application is running) If you are unable to do the above, check the command prompt to fetch the URL of the flask server.