Linked Presentation: Ladder: Enabling Efficient Low-Precision Deep Learning Computing through Hardware-aware Tensor Transformation
