C3W1 - Question about training


I am working on Classify Images of Clouds in the Cloud with AutoML Vision assignment. I have a question.
When we hit “train”, which model is adapted? Have we chose the model?

Where did we choose the search space and strategy then?


1 Like

Hi Ammar! As mentioned by Robert in the lectures, the low-level details are usually not exposed in the AutoML services of Cloud providers. These offerings are meant for automating the entire ML pipeline so it allows even non-experts to use machine learning. The models are chosen automatically based on the chosen application. However, as Robert also said, the algorithms used under the hood to tune and train these models will be similar to the concepts discussed in class. Hope this helps!

1 Like