In this tutorial we will do training of our custom detector . we created the dataset and we annotate them . Next we downloaded the pre -train weight and we started from scratch and did forward and backward pass in convolution network to achieve a lower loss around 0.068 avg.
like and subscribe for next update.