./dma_bin
	Load fixed point weights, convert to short and feed them into SDRAM

./fixed_point
	The set of tools in python to implement fixed-point conversion / operations

./img
	The code to wash data, i.e. resize and clean the original image to fit the requirements of VGG-16 

./VGG_C_implementation
	The complete C implementation of the neural network, we verified that it works well

./vgg16_keras
	The pre-trained we get. We made some adjustment so it can help us to verify other programs

./vgg16_python
	The Winograd and vanilla implementation of VGG-16 network 