CNN-RGB-parallel-processing

require 'torch'

torch.setdefaulttensortype('torch.FloatTensor') torch.setnumthreads(4)

DATA_DIR="./data"

CLASSES = {} LABEL2ID = { frog = 1, truck = 2, deer = 3, automobile = 4, bird = 5, horse = 6, ship = 7, cat = 8, dog = 9, airplane = 10 } ID2LABEL = {} for k, v in pairs(LABEL2ID) do ID2LABEL[v] = k CLASSES[v] = k end

return true