trainingFromScratch

Train from scratch

By using our code, we reproduce the results of our Fast MPN-COV ResNet models on ImageNet 2012.

Our experiments are running on

• PyTorch 0.4 or above
• 2 x 1080Ti
• Cuda 9.0 with CuDNN 7.0

Results

Classification results (single crop 224x224, %) on ImageNet 2012 validation set

Network	Top-1 Error		Top-5 Error
	paper	reproduce	paper	reproduce
fast MPN-COV-ResNet50	22.14	21.71	6.22	6.13
fast MPN-COV-ResNet101	21.21	20.99	5.68	5.56