You can now use images not used in the training the model to infer and classify your species of interest, here ensure that the model has not previously seen this image in any way. This can be achieved by using the built-in prediction tools in fast.ai as illustrated in the model inference section code of the development phase on the Jupyter Notebook.