RpSort16
产品支持情况
产品 |
是否支持 |
|---|---|
Atlas 350 加速卡 |
x |
x |
|
x |
|
x |
|
√ |
|
x |
|
√ |
功能说明
根据Region Proposals中的score域对其进行排序（score大的排前面），每次排16个Region Proposals。
函数原型
1 2 | template <typename T> __aicore__ inline void RpSort16(const LocalTensor<T>& dst, const LocalTensor<T>& src, const int32_t repeatTime) |
参数说明
参数名 |
描述 |
|---|---|
T |
操作数数据类型。 |
参数名称 |
输入/输出 |
含义 |
|---|---|---|
dst |
输出 |
目的操作数，存储经过排序后的Region Proposals。 类型为LocalTensor，支持的TPosition为VECIN/VECCALC/VECOUT。 LocalTensor的起始地址需要32字节对齐。 |
src |
输入 |
源操作数，存储未经过排序的Region Proposals。 类型为LocalTensor，支持的TPosition为VECIN/VECCALC/VECOUT。 LocalTensor的起始地址需要32字节对齐。 |
repeatTime |
输入 |
重复迭代次数，int32_t类型，每次排16个Region Proposals。取值范围：repeatTime∈[0,255]。 |
约束说明
- 用户需保证src和dst中存储的Region Proposal数目大于实际所需数据，否则会存在tensor越界错误。
- 当存在proposal[i]与proposal[j]的score值相同时，如果i>j，则proposal[j]将首先被选出来，排在前面。
- 操作数地址对齐要求请参见通用地址对齐约束。
调用示例
- 接口使用样例
1 2 3 4
// ProposalConcat将连续元素合入Region Proposal内对应位置 // repeatTime = 2, 对2个Region Proposal进行排序，model=4起始位置为4 AscendC::ProposalConcat(dstLocal, srcLocal, 2, 4); AscendC::RpSort16(dstLocal, dstLocal, 2);
示例结果 输入数据srcLocal: [ -1.624 -42.3 -54.12 91.25 -99.4 36.72 67.44 -66.3 -52.53 3.377 -62.47 -15.85 -31.47 3.143 58.47 -83.75 21.58 63.47 7.234 35.16 -39.72 37.8 73.06 -98.7 44.1 -77.2 67.2 19.62 -87.9 -14.875 15.86 -77.75] 经过ProposalConcat后的dstLocal数据，repeat=2计算32个元素，model=4起始位置为4 [ 0. 0. 0. 0. -1.624 0. 0. 0. 0. 0. 0. 0. -42.3 0. 0. 0. 0. 0. 0. 0. -54.12 0. 0. 0. 0. 0. 0. 0. 91.25 0. 0. 0. 0. 0. 0. 0. -99.4 0. 0. 0. 0. 0. 0. 0. 36.72 0. 0. 0. 0. 0. 0. 0. 67.44 0. 0. 0. 0. 0. 0. 0. -66.3 0. 0. 0. 0. 0. 0. 0. -52.53 0. 0. 0. 0. 0. 0. 0. 3.377 0. 0. 0. 0. 0. 0. 0. -62.47 0. 0. 0. 0. 0. 0. 0. -15.85 0. 0. 0. 0. 0. 0. 0. -31.47 0. 0. 0. 0. 0. 0. 0. 3.143 0. 0. 0. 0. 0. 0. 0. 58.47 0. 0. 0. 0. 0. 0. 0. -83.75 0. 0. 0. 0. 0. 0. 0. 21.58 0. 0. 0. 0. 0. 0. 0. 63.47 0. 0. 0. 0. 0. 0. 0. 7.234 0. 0. 0. 0. 0. 0. 0. 35.16 0. 0. 0. 0. 0. 0. 0. -39.72 0. 0. 0. 0. 0. 0. 0. 37.8 0. 0. 0. 0. 0. 0. 0. 73.06 0. 0. 0. 0. 0. 0. 0. -98.7 0. 0. 0. 0. 0. 0. 0. 44.1 0. 0. 0. 0. 0. 0. 0. -77.2 0. 0. 0. 0. 0. 0. 0. 67.2 0. 0. 0. 0. 0. 0. 0. 19.62 0. 0. 0. 0. 0. 0. 0. -87.9 0. 0. 0. 0. 0. 0. 0. -14.875 0. 0. 0. 0. 0. 0. 0. 15.86 0. 0. 0. 0. 0. 0. 0. -77.75 0. 0. 0. ] 输出数据(dst_gm): [ 0. 0. 0. 0. 91.25 0. 0. 0. 0. 0. 0. 0. 67.44 0. 0. 0. 0. 0. 0. 0. 58.47 0. 0. 0. 0. 0. 0. 0. 36.72 0. 0. 0. 0. 0. 0. 0. 3.377 0. 0. 0. 0. 0. 0. 0. 3.143 0. 0. 0. 0. 0. 0. 0. -1.624 0. 0. 0. 0. 0. 0. 0. -15.85 0. 0. 0. 0. 0. 0. 0. -31.47 0. 0. 0. 0. 0. 0. 0. -42.3 0. 0. 0. 0. 0. 0. 0. -52.53 0. 0. 0. 0. 0. 0. 0. -54.12 0. 0. 0. 0. 0. 0. 0. -62.47 0. 0. 0. 0. 0. 0. 0. -66.3 0. 0. 0. 0. 0. 0. 0 -83.75 0. 0. 0. 0. 0. 0. 0. -99.4 0. 0. 0. 0. 0. 0. 0. 73.06 0. 0. 0. 0. 0. 0. 0. 67.2 0. 0. 0. 0. 0. 0. 0. 63.47 0. 0. 0. 0. 0. 0. 0. 44.1 0. 0. 0. 0. 0. 0. 0. 37.8 0. 0. 0. 0. 0. 0. 0. 35.16 0. 0. 0. 0. 0. 0. 0. 21.58 0. 0. 0. 0. 0. 0. 0. 19.62 0. 0. 0. 0. 0. 0. 0. 15.86 0. 0. 0. 0. 0. 0. 0. 7.234 0. 0. 0. 0. 0. 0. 0. -14.875 0. 0. 0. 0. 0. 0. 0. -39.72 0. 0. 0. 0. 0. 0. 0. -77.2 0. 0. 0. 0. 0. 0. 0. -77.75 0. 0. 0. 0. 0. 0. 0. -87.9 0. 0. 0. 0. 0. 0. 0. -98.7 0. 0. 0. ]
父主题： 排序组合（ISASI）