https://github.com/amscotti/local-llm-with-rag
Running local Language Language Models (LLM) to perform Retrieval-Augmented Generation (RAG)
https://github.com/amscotti/local-llm-with-rag
chatbot embeddings langchain llm mistral ollama python rag retrieval-augmented-generation streamlit
Last synced: 3 months ago
JSON representation
Running local Language Language Models (LLM) to perform Retrieval-Augmented Generation (RAG)
- Host: GitHub
- URL: https://github.com/amscotti/local-llm-with-rag
- Owner: amscotti
- License: mit
- Created: 2023-11-05T21:35:34.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-02-12T23:09:39.000Z (5 months ago)
- Last Synced: 2025-04-04T09:06:43.508Z (3 months ago)
- Topics: chatbot, embeddings, langchain, llm, mistral, ollama, python, rag, retrieval-augmented-generation, streamlit
- Language: Python
- Homepage:
- Size: 18.2 MB
- Stars: 220
- Watchers: 7
- Forks: 41
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- License: LICENSE
Awesome Lists containing this project
README
# Local LLM with RAG
This project is an experimental sandbox for testing out ideas related to running local Large Language Models (LLMs) with [Ollama](https://ollama.ai/) to perform Retrieval-Augmented Generation (RAG) for answering questions based on sample PDFs. In this project, we are also using Ollama to create embeddings with the [nomic-embed-text](https://ollama.com/library/nomic-embed-text) to use with [Chroma](https://docs.trychroma.com/). Please note that the embeddings are reloaded each time the application runs, which is not efficient and is only done here for testing purposes.
[](https://asciinema.org/a/fepTvXf1UiDpRUhhNiswL8isu)
There is also a web UI created using [Streamlit](https://streamlit.io/) to provide a different way to interact with Ollama.
## Requirements
- [Ollama](https://ollama.ai/) verson 0.5.7 or higher.
## Setup
1. Clone this repository to your local machine.
2. Install UV using instructions from the Astral site, [Installation](https://docs.astral.sh/uv/#installation)
3. Create a virtual environment and install the required Python packages by running `uv sync`
## Running the Project
**Note:** The first time you run the project, it will download the necessary models from Ollama for the LLM and embeddings. This is a one-time setup process and may take some time depending on your internet connection.
1. Run the main script with `uv app.py -m -p ` to specify a model and the path to documents. If no model is specified, it defaults to [mistral](https://ollama.com/library/mistral). If no path is specified, it defaults to `Research` located in the repository for example purposes.
2. Optionally, you can specify the embedding model to use with `-e `. If not specified, it defaults to [nomic-embed-text](https://ollama.com/library/nomic-embed-text).
This will load the PDFs and Markdown files, generate embeddings, query the collection, and answer the question defined in `app.py`.
## Running the Streamlit UI
Run the Streamlit application by executing `uv streamlit run ui.py` in your terminal.
This will start a local web server and open a new tab in your default web browser where you can interact with the application. The Streamlit UI allows you to select models, select a folder, providing an easier and more intuitive way to interact with the RAG chatbot system compared to the command-line interface. The application will handle the loading of documents, generating embeddings, querying the collection, and displaying the results interactively.
## Technologies Used
- [Langchain](https://github.com/langchain/langchain): A Python library for working with Large Language Model
- [Ollama](https://ollama.ai/): A platform for running Large Language models locally.
- [Chroma](https://docs.trychroma.com/): A vector database for storing and retrieving embeddings.
- [PyPDF](https://pypi.org/project/PyPDF2/): A Python library for reading and manipulating PDF files.
- [Streamlit](https://streamlit.io/): A web framework for creating interactive applications for machine learning and data science projects.
- [UV](https://astral.sh/uv): A fast and efficient Python package installer and resolver.