How can I determine the optimal number of layers and units for a neural network, and what factors should I consider during this process?
Hi @ThisisAva working with neural network requires a lot trial and error, there is no practical way to know the right number, it depends on your data and your trials
1 Like
There is also a way to go about and do it I mean search for an optimal structure and hyper-parameters as well automating the process in some way. As far as I remember the MLops Specialization introduces some of these (I am not sure if its accessible now) and some other courses here too, check the Deep Learning Specialization (read through the courses content).
Here is an article to give you some insight but do check courses here because the way they are explained here is very easy to understand, this is my personal observation!