-- high level training of EfficentDet Lite -- made by google
-- Uses TFOD2 (stable) to train effdetD6 as well as run inference
-- Given a checkpoint, builds frozen graph that can be exported to tflite
-- Attempt to perform full integer quantization on effDet D6 -- TODO: train D0 and try that, EdgeTPU compile converted D6 file to determine feasibility
-- Masks_RoboTrain_Inference