Developing real-world applications

Question 1

Factors to consider when approaching a new problem

Answer 1

• The number of targets/how many classes?
• How vast is the intra-class variance?
• How vast is the inter-class variance?
• How big is the data?
• How balanced is the data?
• Is there already a model that is trained with a lot of images?

Question 2

Choosing the right model

Answer 2

There is a trade-off between size, runtime and accuracy. Test out all the architectures before the final decision.