FTorch and TorchFort

These are complementary tools that serve different integration patterns: FTorch enables Fortran code to invoke PyTorch models as external inference engines, while TorchFort provides an online learning interface for coupling neural networks directly into HPC simulation loops on GPUs.

FTorch
57
Established
TorchFort
56
Established
Maintenance 10/25
Adoption 10/25
Maturity 16/25
Community 21/25
Maintenance 10/25
Adoption 10/25
Maturity 16/25
Community 20/25
Stars: 185
Forks: 39
Downloads:
Commits (30d): 0
Language: Fortran
License: MIT
Stars: 184
Forks: 33
Downloads:
Commits (30d): 0
Language: C++
License: Apache-2.0
No Package No Dependents
No Package No Dependents

About FTorch

Cambridge-ICCS/FTorch

A library for directly calling PyTorch ML models from Fortran.

This library helps scientists and engineers who use Fortran for high-performance computing integrate modern machine learning models into their existing simulations. You can take a PyTorch model trained in Python, feed it into your Fortran code, and get predictions back, all within the same Fortran application. This is ideal for researchers in fields like physics, climate modeling, or aerospace where Fortran is prevalent.

scientific-computing numerical-simulation high-performance-computing physics-modeling engineering-analysis

About TorchFort

NVIDIA/TorchFort

An Online Deep Learning Interface for HPC programs on NVIDIA GPUs

This tool helps scientists and engineers integrate Deep Learning directly into their high-performance computing (HPC) simulation codes. You can feed data arrays from your Fortran or C/C++ simulations into PyTorch models for training or inference, all within the same simulation process. This is ideal for domain scientists working with complex simulations who want to embed machine learning capabilities.

scientific-simulation computational-science high-performance-computing physics-modeling engineering-simulation

Scores updated daily from GitHub, PyPI, and npm data. How scores work