PhysicsNeMo
PhysicsNeMo is NVIDIA's open-source physics-ML framework (v2.x in 2026, Apache-2.0) offering complete training from data-driven to physics-constrained (PINN / soft & hard constraints), with preset physics equations and distributed training. Its Apache-2.0 license is very friendly to closed-source commercial use, making it the mainstream framework for industrial physics AI.
Use cases: Training industrial surrogates; embedding physics-constrained networks; large-scale training with HPC / GPU clusters.
Tags: PINNNeural OperatorNVIDIACommercial-Friendly
Related entries
- AIPerf: LLM Inference Benchmarking at Scale · AIPerf is an LLM inference benchmarking tool from NVIDIA designed to evaluate th
- How NVIDIA NVLink 6 Delivers Multi-Layer Resiliency for AI Factories · NVIDIA NVLink 6 introduces multi-layer resiliency mechanisms designed to enhance
- Accelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine · NVIDIA Transformer Engine introduces JAX-optimized kernels for Mixture of Expert
- PhysicsNeMo v2.2.2 · NVIDIA PhysicsNeMo v2.2.2 release fixes documentation alignment with the PyPi pa
- DeepXDE · The most beginner-friendly PINN framework, with excellent Chinese docs and Tenso