TY - GEN
T1 - Adaptive Tiling
T2 - 24th International Conference on Pattern Recognition, ICPR 2018
AU - Kung, H. T.
AU - McDanel, Bradley
AU - Zhang, Sai Qian
N1 - Publisher Copyright:
© 2018 IEEE.
PY - 2018/11/26
Y1 - 2018/11/26
N2 - We introduce adaptive tiling, a method of partitioning layers in a sparse convolutional neural network (CNN) into blocks of filters and channels, called tiles, each implementable with a fixed-size systolic array. By allowing a tile to adapt its size so that it can cover a large sparse area, we minimize the total number of tiles, or equivalently, the number of systolic array calls required to perform CNN inference. The proposed scheme resolves a challenge of applying systolic array architectures, traditionally designed for dense matrices, to sparse CNNs. To validate the approach, we construct a highly sparse Lasso-Mobile network by pruning MobileNet trained with an ℓ1 regularization penalty, and demonstrate that adaptive tiling can lead to a 2- 3x reduction in systolic array calls, on Lasso-Mobile, for several benchmark datasets.
AB - We introduce adaptive tiling, a method of partitioning layers in a sparse convolutional neural network (CNN) into blocks of filters and channels, called tiles, each implementable with a fixed-size systolic array. By allowing a tile to adapt its size so that it can cover a large sparse area, we minimize the total number of tiles, or equivalently, the number of systolic array calls required to perform CNN inference. The proposed scheme resolves a challenge of applying systolic array architectures, traditionally designed for dense matrices, to sparse CNNs. To validate the approach, we construct a highly sparse Lasso-Mobile network by pruning MobileNet trained with an ℓ1 regularization penalty, and demonstrate that adaptive tiling can lead to a 2- 3x reduction in systolic array calls, on Lasso-Mobile, for several benchmark datasets.
UR - http://www.scopus.com/inward/record.url?scp=85059757793&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85059757793&partnerID=8YFLogxK
U2 - 10.1109/ICPR.2018.8545462
DO - 10.1109/ICPR.2018.8545462
M3 - Conference contribution
AN - SCOPUS:85059757793
T3 - Proceedings - International Conference on Pattern Recognition
SP - 1006
EP - 1011
BT - 2018 24th International Conference on Pattern Recognition, ICPR 2018
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 20 August 2018 through 24 August 2018
ER -