modulus.discrete

discrete.constraints.constraint

Continuous type contraints

class modulus.discrete.constraints.constraint.DeepONetConstraint_Data(nodes: List[modulus.node.Node], invar_branch: Dict[str, numpy.array], invar_trunk: Dict[str, numpy.array], outvar: Dict[str, numpy.array], batch_size: int, cell_volumes: Optional[numpy.array] = None, lambda_weighting: Optional[Dict[str, Union[numpy.array, sympy.Basic]]] = None, lazy_loading: bool = False, num_workers: int = 0, loss=PointwiseLossNorm())

Bases: modulus.constraint.Constraint

class modulus.discrete.constraints.constraint.DeepONetConstraint_Physics(nodes: List[modulus.node.Node], invar_branch: Dict[str, numpy.array], invar_trunk: Dict[str, numpy.array], outvar: Dict[str, numpy.array], batch_size: int, tile_trunk_input: bool = True, cell_volumes: Optional[numpy.array] = None, lambda_weighting: Optional[Dict[str, Union[numpy.array, sympy.Basic]]] = None, lazy_loading: bool = False, num_workers: int = 0, loss=PointwiseLossNorm())

Bases: modulus.constraint.Constraint

class modulus.discrete.constraints.constraint.GridConstraint(dataset, nodes, num_workers, loss=PointwiseLossNorm())

Bases: modulus.constraint.Constraint

Grid Constraint

class modulus.discrete.constraints.constraint.SupervisedGridConstraint(nodes: List[modulus.node.Node], invar: Union[Dict[str, numpy.array], Dict[str, modulus.discrete.dataset.datafile.DataFile]], outvar: Union[Dict[str, numpy.array], Dict[str, modulus.discrete.dataset.datafile.DataFile]], batch_size: int, cell_volumes: Optional[numpy.array] = None, lambda_weighting: Optional[Dict[str, Union[numpy.array, sympy.Basic]]] = None, lazy_loading: bool = False, num_workers: int = 0, loss: modulus.loss.Loss = PointwiseLossNorm())

Bases: modulus.discrete.constraints.constraint.GridConstraint

Data-driven grid field constraint

nodesList[Node]

List of Modulus Nodes to unroll graph with.

invarUnion[Dict[str, np.array], Dict[str, DataFile]]

Dictionary of numpy arrays as input. Input arrays should be of form [B, cin, xdim, …] If lazy loading is being used this should be a dictionary of DataFile objets

outvarUnion[Dict[str, np.array], Dict[str, DataFile]]

Dictionary of numpy arrays as target outputs. Target arrays should be of form [B, cin, xdim, …] If lazy loading is being used this should be a dictionary of DataFile objets

batch_sizeint

Batch size used when running constraint.

lambda_weightingDict[str, Union[np.array, sp.Basic]], optional

The weighting of the constraint, by default None

lazy_loadingbool, optional

Use lazy-loading of dataset, by default False

num_workersint, optional

Number of dataloader workers, by default 0

lossLoss, optional

Modulus Loss function, by default PointwiseLossNorm()

discrete.dataset.datafile

class modulus.discrete.dataset.datafile.DataFile(cache_size: Optional[int] = None)

Bases: object

Base class for storing datafile information needed for lazy loading data. Should typically be used with large data-sets that consist of multiple files or can be partially loaded (e.g. HDF5).

class modulus.discrete.dataset.datafile.HDF5DataFile(file_name_or_path: Union[str, pathlib.Path], variable_name: str, cache_size: Optional[int] = None)

Bases: modulus.discrete.dataset.datafile.DataFile

Datafile class for a HDF5 file

discrete.dataset.dataset

Modulus Dataset constructors for continuous type data

class modulus.discrete.dataset.dataset.GridDataset(*args, **kwds)

Bases: torch.utils.data.dataset.IterableDataset, modulus.discrete.dataset.dataset.DiscreteDataset

Base class for data-driven learning of data on stuctured grid

class modulus.discrete.dataset.dataset.SupervisedGridDataset(batch_size: int, invar: Union[Dict[str, numpy.array], Dict[str, modulus.discrete.dataset.datafile.DataFile]], outvar: Union[Dict[str, numpy.array], Dict[str, modulus.discrete.dataset.datafile.DataFile]], lambda_weighting: Optional[Dict[str, Union[numpy.array, sympy.Basic]]] = None, shuffle: bool = True, drop_last: bool = False, lazy_loading: bool = True)

Bases: modulus.discrete.dataset.dataset.GridDataset

An infitely iterable dataset for data-driven learning of data on stuctured grid

batch_sizeint

Batch size of dataset

invarUnion[Dict[str, np.array], Dict[str, DataFile]]

Dictionary of numpy arrays as input. Input arrays should be of form [B, cin, xdim, …] If lazy loading is being used this should be a dictionary of DataFile objets

outvarUnion[Dict[str, np.array], Dict[str, DataFile]]

Dictionary of numpy arrays as target outputs. Target arrays should be of form [B, cin, xdim, …] If lazy loading is being used this should be a dictionary of DataFile objets

lambda_weightingDict[str, Union[np.array, sp.Basic]], optional

The weighting of the each example, by default None

shufflebool, optional

Randomly shuffle examples in dataset every epoch, by default True

drop_lastbool, optional

Drop last mini-batch if dataset not fully divisible but batch_size, by default False

lazy_loadingbool, optional

Use lazy-loading, by default True

worker_init_fn(i)

Called by worker process when it initialises in torch DataLoader

class modulus.discrete.dataset.dataset.ValidationGridDataset(batch_size: int, invar: Dict[str, numpy.array], outvar: Dict[str, numpy.array], lambda_weighting: Optional[Dict[str, numpy.array]] = None, shuffle: bool = False, drop_last: bool = False)

Bases: modulus.discrete.dataset.dataset.GridDataset

An finitely iterable dataset for data-driven learning of data on stuctured grid. Designed to be used with validators.

batch_sizeint

Batch size of dataset

invarDict[str, np.array]

Dictionary of numpy arrays as input. Input arrays should be of form [B, cin, xdim, …] If lazy loading is being used this should be a dictionary of DataFile objets

outvarDict[str, np.array]

Dictionary of numpy arrays as target outputs. Target arrays should be of form [B, cin, xdim, …] If lazy loading is being used this should be a dictionary of DataFile objets

lambda_weightingDict[str, np.array], optional

The weighting of the each example, by default None

shufflebool, optional

Randomly shuffle examples in dataset every epoch, by default True

drop_lastbool, optional

Drop last mini-batch if dataset not fully divisible but batch_size, by default False

discrete.solvers.solver

Modulus Neural Differential Equation Solver

class modulus.discrete.solvers.solver.Solver(cfg: omegaconf.dictconfig.DictConfig, domain: modulus.continuous.domain.domain.Domain)

Bases: modulus.continuous.solvers.solver.Solver

Trains and Evaluates solver.

discrete.validator.validator

Validator for Solver class

class modulus.discrete.validator.validator.DeepONet_Data_Validator(invar_branch: Dict[str, numpy.array], invar_trunk: Dict[str, numpy.array], true_outvar: Dict[str, numpy.array], nodes, batch_size: int = 100, plotter=None, requires_grad: bool = True)

Bases: modulus.discrete.validator.validator.Validator

DeepONet Validator

forward_grad(invar)

forward_nograd(invar)

save_results(name, results_dir, writer, save_filetypes, step)

class modulus.discrete.validator.validator.DeepONet_Physics_Validator(invar_branch: Dict[str, numpy.array], invar_trunk: Dict[str, numpy.array], true_outvar: Dict[str, numpy.array], nodes, batch_size: int = 100, plotter=None, requires_grad: bool = True, tile_trunk_input: bool = True)

Bases: modulus.discrete.validator.validator.Validator

DeepONet Validator

forward_grad(invar)

forward_nograd(invar)

save_results(name, results_dir, writer, save_filetypes, step)

class modulus.discrete.validator.validator.GridValidator(invar: Dict[str, numpy.array], true_outvar: Dict[str, numpy.array], nodes: List[modulus.node.Node], batch_size: int = 100, plotter=None, requires_grad: bool = True)

Bases: modulus.discrete.validator.validator.Validator

Data-driven grid field validator

invarDict[str, np.array]

Dictionary of numpy arrays as input. Input arrays should be of form [B, cin, xdim, …]

true_outvarDict[str, np.array]

Dictionary of numpy arrays as target outputs. Target arrays should be of form [B, cin, xdim, …]

nodesList[Node]

List of Modulus Nodes to unroll graph with.

batch_sizeint, optional

Batch size used when running validation, by default 100

plotterUnion[Plotter, None]

Modulus Plotter for showing results in tensorboard.

requires_gradbool = True

If automatic differentiation is needed for computing results.

forward_grad(invar)

forward_nograd(invar)

save_results(name, results_dir, writer, save_filetypes, step)

class modulus.discrete.validator.validator.Validator

Bases: object

Validator base class

save_results(filename)