Mastering Computer Vision with TensorFlow 2.x

Adjusting the YOLO configuration file

After completing these steps, the file arrangement part is complete and we will now work on optimizing the parameters in the YOLO configuration file. To do that, open YOLO-VOC.cfg under the Darknet cfg directory and make the following changes. The resulting code can also be downloaded from https://github.com/PacktPublishing/Mastering-Computer-Vision-with-TensorFlow-2.0/blob/master/Chapter07/yolov3-furniture.cfg:

Note that in the following section, we will be describing various line numbers and values to change – these line numbers correspond to the YOLO-VOC.cfg file.

1. Line 6—the batch size. Set this to 64. This means 64 images will be used in every training step to update the CNN parameters.
2. Line 7—