README.md

Local 🏠 Private 🔒 தமிழ் AI Assistant 🤖

Powered by தமிழ் 🦙🦙🦙

Lets create a fully private, local & offline chat bot using tamil LLM. This repo is accompanying concepts in tamil YT video :

References 🙏🏻

This work is based on Abinand's Tamil-LLAMA work that is published in HuggingFace. It specifically uses instruct-7b-v0.2 version of the model.

Contributions 💁🏻‍♂️

I ported the GGUF model to Ollama format and uploded it to ollama model repo under conceptsintamil/tamil-llama-7b-instruct-v0.2. Now after installing Ollama, simply do the following:

1. run the model directly ollama run conceptsintamil/tamil-llama-7b-instruct-v0.2
2. pull the model into local repo using ollama pull conceptsintamil/tamil-llama-7b-instruct-v0.2

Steps to follow ✅

1. Download and install Ollama by following instructions according to your OS (only linux or mac supported)

2. Download Tamil-LLAMA from the Ollama model repo

ollama pull conceptsintamil/tamil-llama-7b-instruct-v0.2

3. Install pip requirements

pip install streamlit
pip install langchain

4. Run the bot 🏃🏻‍♂️

streamlit run app.py

Now browse the locally brewed ☕️, completely private 🔒, offline working 🙅🏻‍♂️, Tamil Based Bot 🤖