Species Classification Model Training

Datastore

The first step in the training process is to create an Azure Machine Learning datastore. This is done by going to the Azure Machine Learning workspace, and clicking on the Datastores tab. From here, you can create a new datastore, and link it to the Azure Blob Storage container that contains the training data. This is done by selecting Azure Blob Storage as the datastore type, and then entering the name of the container, and the account key. The account key can be found in the Azure Portal, by going to the storage account, and selecting Access Keys from the menu on the left. The account key is one of the two keys listed here.

Data assets

Once the datastore has been created, we can register the data assets. This is done by going to the Datasets tab, and clicking on Create dataset. From here, we can select the datastore that we created in the previous step, and then select the folder that contains the training data. We can then select the type of dataset that we want to create. In this case, we want to create a File dataset, as we are using image files. We can then give the dataset a name, and click Next. We can then select the file types that we want to include in the dataset. In this case, we want to include .jpg and .jpeg files. We can then click Next, and then Create to create the dataset.