SergeyLev

Labels are encoded in image file name eg. 京PL3N67.jpg. In case if you have a numberplate with same symbols, filename should be something like 京PL3N67_bla.jpg

If you want to train for more than 15 epochs, change parameter --max_epoch. If you want to continue training using pre-trained weights, set path in parameter --pretrained_model

Replace yolov4 with what you want -> adjust feature creation.

> > At iteration 1000? See: #8669 > > how to downgrade CUDNN in colab? > > i'm running this code in colab `!pip install libcudnn8-dev==8.4.1.50-1+cuda11.6 libcudnn8==8.4.1.50-1+cuda11.6` > > and...