PhyGHT mu200

Checkpoint Release of PhyGHT mu200. Training completed all 200 epochs on 8,000 events. The checkpoint was selected at epoch 190 using validation hybrid loss only. Test results were measured after training and selection.

Split Target R2 MAE RMSE MSE
validation energy_fraction 0.93059360 0.01236418 0.02045338 0.00041834
validation mass_fraction 0.82539444 0.00956031 0.01575836 0.00024833
test energy_fraction 0.93459781 0.01239546 0.02036713 0.00041482
test mass_fraction 0.83663009 0.00949789 0.01550109 0.00024028

Full results and provenance are in metrics.json, config.json, and source_provenance.json. Complete training history, the training log, and the environment record are included.

Install and infer

The verified runtime is Python 3.10.18 with CUDA-enabled PyTorch 2.6.0+cu124. Install locally in an isolated environment:

python -m pip install torch==2.6.0 --index-url https://download.pytorch.org/whl/cu124
python -m pip install torch-geometric==2.6.1 numpy==1.26.4 safetensors huggingface-hub
python -m pip install --only-binary=:all: 'torch-cluster==1.6.3+pt26cu124' 'torch-scatter==2.1.2+pt26cu124' -f https://data.pyg.org/whl/torch-2.6.0+cu124.html
import json, sys, os
os.environ.setdefault("CUBLAS_WORKSPACE_CONFIG", ":4096:8")
from pathlib import Path
import torch
from huggingface_hub import snapshot_download
folder = Path(snapshot_download("Rakib/PhyGHT-mu200"))
sys.path.insert(0, str(folder))
from portable_io import load_model, predict_event
device = "cuda" if torch.cuda.is_available() else "cpu"
model = load_model(folder, device=device)
event = json.loads((folder / "example_event.json").read_text())
predictions = predict_event(model, event, device=device)  # [number of jets, 2]

Inputs and preprocessing

One event has jets (float64 rows [pT, eta, phi, mass, Efrac, Mfrac]), trks (float64 rows [pT, eta, phi, mass, d0, z0, track_id, truth_label]), and jet_trk_IDs (variable-length lists of int64 identifiers, one per jet). Model features use only the first six track and first four jet columns, cast to float32 without normalization or truncation. Track ID column 6 identifies constituents; identifiers are not track positions. Truth labels and regression targets are excluded from features. At inference, jet rows may contain only four feature columns and track rows seven columns, including track ID.

The auxiliary signal target is truth_label == -1. Training uses energy MSE + mass MSE + 0.1 times track BCE. The model has dimension 128, four heads, three Transformer layers, eight neighbors, dropout 0.1, and AdamW with learning rate 0.0003, weight decay 0.01, and gradient clipping 1.0. Effective batch size is 16. Smaller memory microbatches, if needed, retain per-jet MSE and per-track BCE weighting before one clipped optimizer update per logical batch.

Events use the exact seed-42 permutation of all 10,000 original events, split 8,000/1,000/1,000. No new events were simulated and no dataset files were changed on the Hub. Use mu200 and mu60 as separate configurations.

Attribution, license, and limitations

Authors: Mohammed Rakib, Luke Vaughan, Shivang Patel, Flera Rizatdinova, Alexander Khanov, and Atriya Sen.

The included source code is MIT licensed (CODE_LICENSE).

Downloads last month
42
Safetensors
Model size
954k params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Dataset used to train Rakib/PhyGHT-mu200

Paper for Rakib/PhyGHT-mu200