Python helper functions for Darknet usage
Plots 2-D learning curve (abscissa=epochs, ordinate=AVG training loss) given a file named './train.out'. train.out is the stdout of Darknet training
./darknet detector train <data> <cfg> <arch> > train.out
Detects standard file './predictions.png' and moves file to './result/predictions-x.png' with increasing index x >=0. It allows processing of multiple images referred by a text file './imgfiles.txt'. Example:
cat imgfiles.txt
data/horses.jpg
data/dog.jpg
./cpproc.sh &
./darknet detect cfg/yolo.cfg yolo.weights < imgfiles.txt
ls result
predictions-0.png predictions-1.png
pgrep -f cpproc.sh | xargs kill -15