Welcome to the LLaMA Math Wizard project! We're teaching an AI llama to be a math genius. 🎓

Ever wished you had a magical llama that could solve math problems? Well, now you do! We've taken the brilliant LLaMA-2-7b-chat model and sent it to math school. The result? A fuzzy, four-legged, formula-crunching friend!

• 🧠 Fine-tuned LLaMA-2-7b-chat model with a PhD in Mathematics
• 📊 Trained on a diverse mathematics dataset
• 🖥️ Streamlit app for chatting with your new math tutor
• 📈 TensorBoard integration for beautiful learning curves
• 🔍 QLoRA magic for efficient training

1. Clone this repository (no llamas were harmed in the process):

git clone https://github.com/utkarshpophli/llama-math-finetuning.git
cd llama-math-wizard

2. Install the required packages (llama food):

pip install -r requirements.txt

3. Log in to Hugging Face (llama face optional):

huggingface-cli login

1. Train your llama:

python src/train.py

2. Chat with your math wizard:

streamlit run app.py

• src/: The llama's brain 🧠
• app/: The llama's mouth 🗣️
• notebooks/: The llama's sandbox 🏖️
• logs/: The llama's diary 📔

Got ideas to make our llama even smarter? We're all ears (and wool)! Feel free to open an issue or submit a pull request.

This project is licensed under the MIT License - see the LICENSE file for details.

• The incredible Hugging Face 🤗 team
• The LAION AI team for blessing us with LLaMA
• Caffeine ☕, for fueling late-night coding sessions

Remember: Be kind to llamas, they're doing their best to understand math! 🦙❤️🔢