I am training Chainer SSD300 model in GeForce 2080 RTX Ti.
The following scenario is for same dataset, tried the whole range of batch-size(4-32), and epochs 100-250.
