DFODE-kit¶
DFODE-kit is a Python toolkit for accelerating combustion chemistry integration with deep learning. It supports the end-to-end workflow of:
- sampling thermochemical states from canonical flame cases,
- augmenting and labeling datasets,
- training neural network surrogates for stiff chemistry integration,
- validating and deploying them in DeepFlame-based CFD workflows.
What this docs site covers¶
- Getting Started: environment setup and installation
- CLI: current
dfode-kitcommands 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¶
- GitHub: deepflame-ai/DFODE-kit
- DeepFlame docs: deepflame.deepmodeling.com