EveNet is a pretrained, multi-task foundation model for event-level collider physics. It comes with a scalable Ray + PyTorch Lightning training pipeline, SLURM-ready multi-GPU infrastructure, and modular YAML configuration, so researchers can quickly fine-tune it on their own datasets and extend it to new physics analyses.
Start directly from pretrained EveNet checkpoints for fine-tuning or inference:
👉 HuggingFace: Avencast/EveNet
Explore the full documentation site for setup options, configuration references, and tutorials:
uw-epe-ml.github.io/EveNet_Public
Improvements are welcome! File an issue or open a pull request for bug fixes, new physics processes, or documentation tweaks. When you add new components or datasets, update the relevant markdown guides so future users can follow along easily.