Llama2-Inference

This repository contains example code for performing inference with Llama 2 using 🤗 Transformers on a supercomputer.

The process consists of the following steps:

1. Create an Apptainer file using the tested llama2_2.def and mamba_llama2_2.yml files.
2. Develop a Python script using the llama2.ipynb Jupyter Notebook. Jupyter Notebooks are a favourite among data scientists!
3. Create a Python as llama2.py. My script is quite a mess as I worked alone. No one will adjust it.
4. Create Slurm scripts 7b_prompt-1.sh and 13b_prompt-1.sh to run the Python script that we created.
5. Submit the Slurm job to execute it.