In this section, we will learn how to use YOLO v3 to train your own custom detector. The training process involves a number of different steps. For the sake of clarity, the input and output from each step are indicated in the following flowchart. Many of the training steps are included in YOLO's YOLOv3: An Incremental Improvement publication by Redmon, Joseph, Farhadi, and Ali, published on arXiv in 2018. They are also included under the Training YOLO on VOC section at https://pjreddie.com/darknet/yolo/.
The following image shows you how to use YOLO v3 to train a VOC dataset. In our case, we will use our own custom furniture data that we used to classify images using Keras in Chapter 6, Visual Search Using Transfer Learning:
A detailed description of sections 1 to 11 is described here.
