Search papers, labs, and topics across Lattice.
1
0
2
1
Achieve comparable or better performance with significantly reduced model size and computational cost by replacing standard convolutional kernels with structured, kernelized higher-order layers.
