cann/ops-cv非连续Tensor说明

非连续的Tensor【免费下载链接】ops-cv本项目是CANN提供的图像处理、目标检测相关的算子库实现网络在NPU上加速计算。项目地址: https://gitcode.com/cann/ops-cv目前大部分算子API的输入aclTensor支持“非连续的Tensor”即一个Tensor可以通过(shape, strides, offset)表示。说明创建aclTensor可通过《算子库接口》中“公共接口aclCreateTensor”实现。示例1例如现有一个shape(6, 5)、strides(10, 1)、offset22的Tensor其内存排布如下a0,0, a0,1, a0,2, a0,3, a0,4, a0,5, a0,6, a0,7, a0,8, a0,9a1,0, a1,1, a1,2, a1,3, a1,4, a1,5, a1,6, a1,7, a1,8, a1,9a2,0, a2,1,a2,2, a2,3, a2,4, a2,5, a2,6, a2,7, a2,8, a2,9a3,0, a3,1,a3,2, a3,3, a3,4, a3,5, a3,6, a3,7, a3,8, a3,9a4,0, a4,1,a4,2, a4,3, a4,4, a4,5, a4,6, a4,7, a4,8, a4,9a5,0, a5,1,a5,2, a5,3, a5,4, a5,5, a5,6, a5,7, a5,8, a5,9a6,0, a6,1,a6,2, a6,3, a6,4, a6,5, a6,6, a6,7, a6,8, a6,9a7,0, a7,1,a7,2, a7,3, a7,4, a7,5, a7,6, a7,7, a7,8, a7,9a8,0, a8,1, a8,2, a8,3, a8,4, a8,5, a8,6, a8,7, a8,8, a8,9a9,0, a9,1, a9,2, a9,3, a9,4, a9,5, a9,6, a9,7, a9,8, a9,9即该Tensor排布如上图的深色位置。这个完整的Tensor在内存排布上是不连续的strides描述Tensor维度上相邻两个元素的间隔如果在维度1上的stride为1 该维度是连续的如果在维度0上的stride为10那么相邻的元素间隔10个元素即非连续。offset表示这个Tensor的首元素相对addr的偏移。示例2例如现有一个shape(4, 3)、strides(20, 2)、offset22的Tensor其内存排布如下a0,0, a0,1, a0,2, a0,3, a0,4, a0,5, a0,6, a0,7, a0,8, a0,9a1,0, a1,1, a1,2, a1,3, a1,4, a1,5, a1,6, a1,7, a1,8, a1,9a2,0, a2,1,a2,2, a2,3,a2,4, a2,5,a2,6, a2,7, a2,8, a2,9a3,0, a3,1, a3,2, a3,3, a3,4, a3,5, a3,6, a3,7, a3,8, a3,9a4,0, a4,1,a4,2, a4,3,a4,4, a4,5,a4,6, a4,7, a4,8, a4,9a5,0, a5,1, a5,2, a5,3, a5,4, a5,5, a5,6, a5,7, a5,8, a5,9a6,0, a6,1,a6,2, a6,3,a6,4, a6,5,a6,6, a6,7, a6,8, a6,9a7,0, a7,1, a7,2, a7,3, a7,4, a7,5, a7,6, a7,7, a7,8, a7,9a8,0, a8,1,a8,2, a8,3,a8,4, a8,5,a8,6, a8,7, a8,8, a8,9a9,0, a9,1, a9,2, a9,3, a9,4, a9,5, a9,6, a9,7, a9,8, a9,9即该Tensor排布如上图的深色位置。这个完整的Tensor在内存排布上是不连续的strides描述Tensor维度上相邻两个元素的间隔如果在维度1上的stride为2 该维度上间隔1个元素如果在维度0上的stride为20那么相邻的元素间隔20个元素即非连续。offset表示这个Tensor的首元素相对addr的偏移。【免费下载链接】ops-cv本项目是CANN提供的图像处理、目标检测相关的算子库实现网络在NPU上加速计算。项目地址: https://gitcode.com/cann/ops-cv创作声明：本文部分内容由AI辅助生成（AIGC），仅供参考