Takes as input a N × C × H × W tensor, and a kernel size (h, w) or k interpreted as (k, k), applies the max-pooling on each channel of each sample.
確定! 回上一頁