This is an implementation of Single Shot MultiBox Detector in PyTorch trained on plant images, done as a part of an academic project, Design Practicum. Below poster represents a gist of the motive behind project.
I do not recommend training SSD from scratch. Using pretrained VGG model helps a lot to achieve lower losses. I use the pretrained pytorch/vision VGG16 model from PyTorch model zoo.
This implementation is initially inspired by: