Implementation for YOLOV3 written in Tensorflow 2.

It's currently a work in progress!

To-Do:

1. Config for kicking off training. a. Pretrained model b. Train Configuration c. Save directory d. Evaluation Script e. Inference Script
2. Loss a. Implement and verify all kinds of losses. b. Implement a config for losses.
3. Loading from pre-trained models from darknet
4. Conversion to TensorRT.

Tasks:

1. Write a train step, verify distribution strategy.
2. Overfit on one image
3. Conversion to TensorRT
4. Eval/Predict step.