The goal is to predict the speed of a car from a video. The network consists of a CNN that uses a stack of 4 images.
All training code is contained within train.py. Simply run:
python main.py
To evaluate a trained network, the network parameters have to be save in the output folder. Then, run:
python eval.py --model_name my_model
It loads a model in output folder and uses a portion of the data in train.mp4.