llama-benchmarks

Utilities to help you experiment with and benchmark Meta's Llama models.

Prerequisites

• Python 3.12
• uv (pip install uv)
• Llama models (Download from Meta https://www.llama.com/)

Build and Test Workflow

Configure Environment

source environment.sh

Build Artifacts

Builds local Python venv, installs dependencies, and downloads MMLU dataset needed for tests.

make

Run Tests

Runs test automation.

make tests

Experiments

• Exploring Order Dependency