I'm trying to object detection with yolov3 and yolov4 and i just want to train 1 class= person. I'm using COCO datasets.
After training how i can tested my weight file? I want to test all "test dataset" not only 1 image.
I have train loss graphic but this is not enough. I want to calculate accuracy, precision, recall etc. I found this examples : https://darknet.gong.im/
./darknet detector test cfg/coco.data cfg/yolov3.cfg weights/yolov3.weights batch ./in_images/ ./out_images/ >./results.txt
Can you tell me how i can compare with success metric this two algorithm?
How can i test yolov3 and yolov4 weight files ?
There is map function for testing the model.
./darknet detector map cfg/coco.data cfg/yolov3.cfg weights/yolov3.weights
If you label your test dataset and give the path of it to the 'valid' field inside the data file, you can use map function over your dataset.
Also you can use -map flag while training to see the map results on the graph. If you use -map flag while training darknet also saves best weight file so you don't need to find the best one.