New experimental modules for topology optimization and automatic differentiation enhanced feature mapping approaches are implemented in emopt/experimental, with corresponding examples in examples/experimental. The AutoDiff methods can result in large improvements in optimization speed for designs with variables that parameterize global geometric features. Please see our preprint below and examples for correct usage. Note: Requires PyTorch installation in your local Python environment. These features are still in development.
New experimental modules for topology optimization and automatic differentiation enhanced feature mapping approaches are implemented in emopt/experimental, with corresponding examples in examples/experimental. The AutoDiff methods can result in large improvements in optimization speed for designs with variables that parameterize global geometric features. Please see our preprint below and examples for correct usage. Note: Requires PyTorch installation in your local Python environment. These features are still in development.