pyRadPlan-dosecalc-Det-VHEE-100-lung

Deterministic VHEE (Very High Energy Electron) dose predictor for lung cases, trained for a single fixed beam energy of 100 MeV (ConvBayes_new architecture, single forward pass). Given a CT cuboid and a beamlet, predicts a local physical dose cuboid.

Usable directly as the AIBeamletEngine dose-calculation engine in pyRadPlan:

pln.prop_dose_calc = {"engine": "AIBeamlet", "model": "pyRadPlan-dosecalc-Det-VHEE-100-lung"}

VHEE plans use one fixed energy per beam. Companion models cover the other supported energies — pyRadPlan-dosecalc-Det-VHEE-150-lung and pyRadPlan-dosecalc-Det-VHEE-200-lung; load the one that matches pln.prop_stf["energy"].

Repository contents

Follows the pyRadPlan ML model contract (pyRadPlan.ml):

File	Purpose
model.py	ConvBayes_new network definition
preprocessor.py	ConvDoseSinglePreprocessor — input assembly + forward pass + output scaling
weights.safetensors	Trained weights
model_config.json	Declarative model/preprocessing/dose-calc configuration

Outputs

• physical_dose — predicted physical dose

Training assumptions

• Radiation mode: VHEE (Very High Energy Electrons)
• Beam energy: 100 MeV (fixed — single-energy model)
• Anatomy: lung
• Trained machine: Generic
• Sampling grid: 2 mm spacing, 52 mm lateral field width (±26 mm), depth window [−350, 350] mm relative to the beamlet origin

Predictions outside this range are not guaranteed to be accurate; AIBeamletEngine warns when the plan falls outside the declared range.

Security note: loading this model executes model.py/preprocessor.py shipped in this repository (gated behind trust_remote_code, default True in pyRadPlan.ml).