Interface

class anemoi.models.interface.AnemoiModelInterface(*, config: DotDict, graph_data: HeteroData, statistics: dict, data_indices: dict, metadata: dict, supporting_arrays: dict = None)

Bases: Module

An interface for Anemoi models.

This class is a wrapper around the Anemoi model that includes pre-processing and post-processing steps. It inherits from the PyTorch Module class.

config

Configuration settings for the model.

Type:: DotDict

id

A unique identifier for the model instance.

Type:: str

multi_step

Whether the model uses multi-step input.

Type:: bool

graph_data

Graph data for the model.

Type:: HeteroData

statistics

Statistics for the data.

Type:: dict

metadata

Metadata for the model.

Type:: dict

supporting_arrays

Numpy arraysto store in the checkpoint.

Type:: dict

data_indices

Indices for the data.

Type:: dict

pre_processors

Pre-processing steps to apply to the data before passing it to the model.

Type:: Processors

post_processors

Post-processing steps to apply to the model’s output.

Type:: Processors

model

The underlying Anemoi model.

Type:: AnemoiModelEncProcDec

predict_step(batch: Tensor, model_comm_group: ProcessGroup | None = None, **kwargs) → Tensor

Prediction step for the model.

Parameters:: batch (torch.Tensor) – Input batched data.
Returns:: Predicted data.
Return type:: torch.Tensor