LLMs for Isotope Production

Author: Logan Endes (github.com/Log45)

Getting Started

To reproduce my work this summer, enter these commands in a terminal of a linux machine (I used Ubuntu):

Install Wget

If you don't have wget on your machine, run these commands: sudo apt-get update cd /tmp apt-get install wget sudo apt-get update

Setup Conda

To install anaconda, use these commands: wget https://repo.anaconda.com/archive/Anaconda3-2023.03-1-Linux-x86_64.sh sha256sum Anaconda3-2023.03-1-Linux-x86_64.sh Navigate through the Anaconda setup, choose an install directory, and then test the connection with the following commands source ~/.bashrc conda info Once conda is working, create a new conda environment using: conda create -n env_name

Configure Environment

Once your environment is created, activate it with: conda activate env_name For this project, I used Python 3.8 to be compatible with galai. conda install python=3.8 Before installing dependencies, find out which version of cuda your graphics card is running using: nvidia-smi Now, install the following dependencies (change pytorch-cuda=11.7 to whatever version of cuda you are running): conda install torchvision torchaudio pytorch-cuda=11.7 git -c pytorch -c nvidia pip install -r requirements.txt pip install git+https://github.com/huggingface/transformers

Run the demo

Currently, the only demo that can be ran is in functions.py (as of writing it checks galactica-standard using a single filter with some html files) python3 functions.py If you are afraid of accidentally killing your connection, run: nohup python3 functions.py