Skip to content

DFODE-kit

DFODE-kit is a Python toolkit for accelerating combustion chemistry integration with deep learning. It supports the end-to-end workflow of:

  1. sampling thermochemical states from canonical flame cases,
  2. augmenting and labeling datasets,
  3. training neural network surrogates for stiff chemistry integration,
  4. validating and deploying them in DeepFlame-based CFD workflows.

What this docs site covers

  • Getting Started: environment setup and installation
  • CLI: current dfode-kit commands and their purpose
  • Canonical Case Initialization: preset-based case setup with preview/apply/config workflows
  • Runtime Configuration and Case Execution: persistent machine-local environment config plus reproducible case launching
  • Data Augmentation: the minimal preset-driven augment CLI, including preview/apply/config round-trip
  • Data Preparation and Training Workflow: the current artifact flow from sampled HDF5 to labeled datasets and models
  • Architecture: repo layout and current refactor direction
  • Tutorials and Workflow: how to think about the DFODE pipeline
  • Agent Docs: operational guidance for coding agents and maintainers

Project goals

The current development direction is to make DFODE-kit:

  • easier for humans and coding agents to use,
  • more reproducible for experiment workflows,
  • more modular for new model architectures and training algorithms,
  • more robust through contracts, tests, and CI.

Repository