Torch implementation of "Multiple Object Recognition with Visual Attention" on Kaggle Cats vs Dogs dataset
note: The code hasn't been cleaned up and this work has been all but abandoned, so it is unlikely to be significantly cleaned up or otherwise improved in the future. The main script is "KCvD_R2_aug.lua". Start there and work out.