Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/jlonge4/local_llama
This repo is to showcase how you can run a model locally and offline, free of OpenAI dependencies.
https://github.com/jlonge4/local_llama
artificial-intelligence langchain llama-cpp llamaindex machinelearning offline python
Last synced: 4 days ago
JSON representation
This repo is to showcase how you can run a model locally and offline, free of OpenAI dependencies.
- Host: GitHub
- URL: https://github.com/jlonge4/local_llama
- Owner: jlonge4
- License: apache-2.0
- Created: 2023-05-20T18:05:24.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-07-12T22:18:50.000Z (4 months ago)
- Last Synced: 2024-08-01T13:35:06.316Z (3 months ago)
- Topics: artificial-intelligence, langchain, llama-cpp, llamaindex, machinelearning, offline, python
- Language: Python
- Homepage:
- Size: 71.3 KB
- Stars: 210
- Watchers: 6
- Forks: 28
- Open Issues: 10
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-ccamel - jlonge4/local_llama - This repo is to showcase how you can run a model locally and offline, free of OpenAI dependencies. (Python)
README
# Local Llama
This project enables you to chat with your PDFs, TXT files, or Docx files entirely offline, free from OpenAI dependencies. It's an evolution of the gpt_chatwithPDF project, now leveraging local LLMs for enhanced privacy and offline functionality.
## Features
- Offline operation: Run in airplane mode
- Local LLM integration: Uses Ollama for improved performance
- Multiple file format support: PDF, TXT, DOCX, MD
- Persistent vector database: Reusable indexed documents
- Streamlit-based user interface
## New Updates
- Ollama integration for significant performance improvements
- Uses nomic-embed-text and llama3:8b models (can be changed to your liking)
- Upgraded to Haystack 2.0
- Persistent Chroma vector database to enable re-use of previously updloaded docs
## Installation
1. Install Ollama from https://ollama.ai/download
2. Clone this repository
3. Install dependencies:
```
pip install -r requirements.txt
```
4. Pull required Ollama models:
```
ollama pull nomic-embed-text
ollama pull llama3:8b
```
## Usage
1. Start the Ollama server:
```
ollama serve
```
2. Run the Streamlit app:
```
python -m streamlit run local_llama_v3.py
```
3. Upload your documents and start chatting!
## How It Works
1. Document Indexing: Uploaded files are processed, split, and embedded using Ollama.
2. Vector Storage: Embeddings are stored in a local Chroma vector database.
3. Query Processing: User queries are embedded and relevant document chunks are retrieved.
4. Response Generation: Ollama generates responses based on the retrieved context and chat history.
## License
This project is licensed under the Apache 2.0 License.
## Acknowledgements
- Ollama team for their excellent local LLM solution
- Haystack for providing the RAG framework
- The-Bloke for the GGUF models