23.6k citations
- University of California, BerkeleyUS235 papers
- California Institute of TechnologyUS22 papers
- Lawrence Berkeley National LaboratoryUS16 papers
- University of ChicagoUS14 papers
- UCLA HealthUS12 papers
- Moscow Institute of Thermal TechnologyRU11 papers
- University of MichiganUS11 papers
- Centre National de la Recherche ScientifiqueFR10 papers
- Google (United States)US10 papers
- University of California, Santa CruzUS10 papers
- Cornell UniversityUS9 papers
- Michigan Science CenterUS9 papers
Showing 2017 · cs.CVShow all
2 papers · 2 filters
cs.CV2017★ 41 cited
Shift: A Zero FLOP, Zero Parameter Alternative to Spatial Convolutions
Bichen Wu, Alvin Wan, Xiangyu Yue +6
Neural networks rely on convolutions to aggregate spatial information. However, spatial convolutions are expensive in terms of model size and computation, both of which grow quadra…
cs.CV2017★ 45 cited
ImageNet Training in Minutes
Yang You, Zhao Zhang, Cho-Jui Hsieh +2
Finishing 90-epoch ImageNet-1k training with ResNet-50 on a NVIDIA M40 GPU takes 14 days. This training requires 10^18 single precision operations in total. On the other hand, the…