Frequently asked questions (FAQs)

Although we have been able to solve this multi-label classification problem, the accuracy we experienced was below par. Therefore, in this section, we will see some frequently asked questions (FAQs) that might already be on your mind. Knowing the answers to these questions might help you to improve the accuracy of the CNNs we trained. Answers to these questions can be found in the Appendix:

1. What are the hyperparameters that I can try tuning while implementing this project?
2. My machine is getting OOP while running this project. What should I do?
3. While training the networks with full images, my GPU is getting OOP. What should I do?
4. I understand that the predictive accuracy using CNN in this project is still very low. Did our network under or overfit? Is there any way to observe how the training went?
5. I am very interested in implementing the same...