In this chapter, we solved an interesting dog versus cat classification problem using the transfer learning technique. We used a pre-trained VGG16 model and its weights, and subsequently we fine-tuned the training with a real-life cat versus dog dataset from Kaggle.
Once the training was complete, we saved the trained model for model persistence and subsequent reuse. We saw that the trained model can successfully detect and differentiate both cat and dog images having very different sizes, qualities, and shapes.
Even the trained model/classifier can be used in solving a real-life cat versus dog problem. The takeaway is that this technique with some minimal effort can be extended and used for solving similar image classification problems, which applies to both binary and multiclass classification problems.
In the next chapter, we will see how to develop an end-to-end project...
