This technical blog series shows developers how to train an accurate, deep learning model using a large public dataset and PyTorch. Then, they can optimize and infer the RetinaNet model with TensorRT and NVIDIA DeepStream.