In the previous segment, we discussed how to implement learning rate decay using callbacks. We also generated augmented images and combined with the original images to be used for training.
Let us now discuss the choice of model evaluation metrics using examples from the medical domain.
In this video, you saw specific use cases of precision and recall. These two measures depend heavily on the use case – even in the field of medicine, both of these can be effectively used.
Let’s now go ahead and make the final run and examine our results.
This brings us to the close of the end-to-end session exploring the flowers dataset. We have built a model with a good AUG at the end of 3 epochs. If you train this using more epochs, you should be able to reach a better AUG value.
In the next session, we will train the model on the chest x-Rays dataset.
Report an error