mindflow.pde.FlowWithLoss

class mindflow.pde.FlowWithLoss(model, loss_fn='mse')[source]

Base class of user-defined data-driven flow prediction problems.

Parameters

model (mindspore.nn.Cell) – A training or test model.
loss_fn (Union[str, Cell]) – Loss function. Default: "mse".

Raises

TypeError – If modle or loss_fn is not mindspore.nn.Cell.
NotImplementedError – If the member function get_loss is not implemented.

Supported Platforms:: Ascend GPU

get_loss(inputs, labels)[source]

Compute the loss of the model.

Parameters

inputs (Tensor) – The input data of model.
labels (Tensor) – True values of the samples.