CodeJeNN: A simple C++ neural network generator for physics applications

Abstract

Machine learning has shown promise for physics applications, but integrating Python-based libraries into high-performance C++ solvers can create performance bottlenecks. We present CodeJeNN, which addresses this gap by auto-generating self-contained C++ code from trained Keras models for inference. This eliminates external dependencies through inlined functions, allowing seamless integration into existing frameworks. We describe the Keras-to-C++ workflow, supported architectures, and limitations. CodeJeNN is demonstrated through inference benchmarks against Keras (eager and optimized) and a computational fluid dynamics (CFD) simulation of a hydrogen-air mixing layer in which viscosity is modeled using a neural network, showing that the generated models integrate directly into a solver while remaining efficient and accurate.

Identification of effective transport physics and optimal feedback control: Experimental demonstration on a 1600-fan wind generator
Tamir Shaqarin, Izz Al-Manasra, TianXiang Zhang +3
arXiv11 Oct 2026PreprintCC BY 4.0Full textFluid dynamicsarXiv
CGC-py: A Monte Carlo Event Generator for Gluon Saturation Physics
Haowu Duan, Cong Yi, Si-Wei Dai +3
arXiv3 Sep 2026PreprintCC0Full textHigh-energy & particle physics Nuclear physicsarXiv
Hierarchical emergence of network bursting in a four-cell central pattern generator model
Krishna Pusuluri, Huiwen Wu, Andrey L. Shilnikov
arXiv12 Sep 2026PreprintCC BY 4.0Full textNeuroscience Statistical & nonlinear physicsarXiv
Time-temperature superposition from multi-temperature chirp rheometry of thermorheologically simple materials via physics-informed Gaussian processes
Isaac Y. Miranda-Valdez, Kourosh Mobredi, Juha Koivisto +1
arXiv23 Oct 2026PreprintCC BY 4.0Full textCondensed matterarXiv
PINNIES: An efficient physics-informed neural network framework for integral operator problems
Alireza Afzal Aghaei, Mahdi Movahedian Moghaddam, Kourosh Parand
Computer Physics Communications6 Jun 2026Peer-reviewedLicence not statedFull text10 citedStatistical & nonlinear physicsOpenAlex

Cite

BibTeX RIS

BibTeX

@misc{arcities2026codejenn,
  title = {{CodeJeNN: A simple C++ neural network generator for physics applications}},
  author = {Jay Arcities and Pavel Popov and Eric J Ching and Kamal Viswanath and Ryan F Johnson},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2607.02746},
  eprint = {2607.02746},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2607.02746},
  note = {CC0},
}

RIS

TY  - UNPB
TI  - CodeJeNN: A simple C++ neural network generator for physics applications
AU  - Jay Arcities
AU  - Pavel Popov
AU  - Eric J Ching
AU  - Kamal Viswanath
AU  - Ryan F Johnson
PY  - 2026
DA  - 2026/10/22
PB  - arXiv
DO  - 10.48550/arxiv.2607.02746
UR  - https://arxiv.org/abs/2607.02746
AB  - Machine learning has shown promise for physics applications, but integrating Python-based libraries into high-performance C++ solvers can create performance bottlenecks. We present CodeJeNN, which addresses this gap by auto-generating self-contained C++ code from trained Keras models for inference. This eliminates external dependencies through inlined functions, allowing seamless integration into existing frameworks. We describe the Keras-to-C++ workflow, supported architectures, and limitations. CodeJeNN is demonstrated through inference benchmarks against Keras (eager and optimized) and a computational fluid dynamics (CFD) simulation of a hydrogen-air mixing layer in which viscosity is modeled using a neural network, showing that the generated models integrate directly into a solver while remaining efficient and accurate.
ER  -

Abstract

Related

Cite