Training your first model

Once Anemoi training is installed, you can run your first model with

anemoi-training train

which will use the default model configurations with missing values.

The training script will intentionally crash as it does not know where your data is stored.

These missing values in the configuration are placeholders for the user to fill in marked with ???. You can find the default configurations in the Anemoi Training repository under src/anemoi/training/config/. Alternatively, the next section will show you how to generate a user config file.

Preparing training components

Anemoi Training requires two primary components to get started:

  1. Graph Definition from Anemoi Graphs: This defines the structure of your machine learning model, including the layers, connections, and operations that will be used during training.

  2. Dataset from Anemoi Datasets : This provides the training data that will be fed into the model. The dataset should be pre-processed and formatted according to the specifications of the Anemoi Datasets module.

In many cases, your organisation may already have some of these prepared, but if not, you can follow the steps below to prepare them.

Step 1: Prepare Your Graph Definition

The first step before training your model is to select or create a graph definition. Anemoi Graphs provides a flexible way to define the architecture of your machine learning model. You can either use a predefined graph that suits common forecasting tasks or customize your own to better fit your specific requirements.

The graph defines the connectivity between your datapoints and the operations that will be performed on them during training. The data flow and “learnable parameters” are defined in the model itself, so the graph is a high-level description of the “connecitivity” within the model architecture.

To prepare your graph:

  • Choose an existing graph definition from Anemoi Graphs or define a new one that matches your forecasting needs.

  • Ensure that the graph is compatible with the input data format and the type of weather forecasting problem you are addressing.

Step 2: Prepare Your Dataset

The next step is to prepare your dataset using Anemoi Datasets. This dataset will be used to train the model defined by your graph.

To prepare your dataset:

  • Select a dataset from Anemoi Datasets or create a new dataset that includes all relevant weather variables and historical data points.

  • Pre-process the data to ensure it is in the correct format and structure required by the graph definition.