Skip to content

dikshantsagar/Neutrino-LLaMa

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Neutrino Event Classification with LLaMA 3.2-11B Vision-Instruct

This repository contains the training and inference code for our paper:
"Adapting Vision-Language Models for Neutrino Event Classification in High-Energy Physics" (submitted to Nature Communications).

We fine-tune the multimodal large language model LLaMA 3.2-11B Vision-Instruct on simulated DUNE Near Detector(LArTPC) pixel maps to classify neutrino interactions into three categories:

  • NuE CC — Electron Neutrino Charged Current interaction
  • NuMu CC — Muon Neutrino Charged Current interaction
  • Neutral Current (NC)

The model takes as input 2D projections (xz, yz) of 3D detector events and learns to recognize distinctive features such as fuzzy electron showers (NuE CC) or long muon tracks (NuMu CC).

Code for running LLama 3.2 Vision and the CNN Baseline are provided in their respective directories, please follow instructions in their respective README.md files.

About

LlaMa 3.2 Vision fine-tuned for Neutrino Event Classification with Explanation Generation

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages