aclnnOneHot
支持的产品型号
- Atlas 推理系列产品。
- Atlas 训练系列产品。
- Atlas A2训练系列产品/Atlas 800I A2推理产品。
接口原型
每个算子分为两段式接口,必须先调用“aclnnOneHotGetWorkspaceSize”接口获取计算所需workspace大小以及包含了算子计算流程的执行器,再调用“aclnnOneHot”接口执行计算。
aclnnStatus aclnnOneHotGetWorkspaceSize(const aclTensor* self, int numClasses, const aclTensor* onValue, const aclTensor* offValue, int64_t axis, aclTensor* out, uint64_t* workspaceSize, aclOpExecutor** executor)
aclnnStatus aclnnOneHot(void* workspace, uint64_t workspaceSize, aclOpExecutor* executor, aclrtStream stream)
功能描述
算子功能:对长度为n的输入self, 经过one_hot的计算后得到一个元素数量为n*k的输出out,其中k的值为numClasses。 输出的元素满足下列公式:
示例:
示例1: self = tensor([0, 1, 2, 0, 1]) numClasses = 5 onValue = tensor([1]) offValue = tensor([0]) axis=-1 out的shape为(5,5) out = tensor([[1, 0, 0, 0, 0], [0, 1, 0, 0, 0], [0, 0, 1, 0, 0], [1, 0, 0, 0, 0], [0, 1, 0, 0, 0]]) 示例2: self = tensor([0, 1, 2, 0, 1]) numClasses = 1 onValue = tensor([1]) offValue = tensor([0]) axis=-1 out的shape为(5,1) out = tensor([[1], [0], [0], [1], [0]]) 示例3: self = tensor([0, 1, 2, 0, 1]) numClasses = 0 onValue = tensor([1]) offValue = tensor([0]) axis=-1 out的shape为(5,0) out = tensor([]) 示例4: self = tensor([[1,2,3]]) # shape (1,3) numClasses = 4 onValue = tensor([1]) offValue = tensor([0]) axis=1 out的shape为(1,4,3) out = tensor([[[0. 0. 0.] [1. 0. 0.] [0. 1. 0.] [0. 0. 1.]]]) # shape (1, 4, 3)
aclnnOneHotGetWorkspaceSize
参数说明:
- self(aclTensor*,计算输入):公式中的self,Device侧的aclTensor,数据类型支持INT32、INT64,支持非连续的Tensor,数据格式支持ND。
- numClasses(int,计算输入):表示类别数,数据类型必须输入INT64。当self为空Tensor时,numClasses的值需大于0;当self不为空Tensor时。numClasses需大于等于0。若numClasses的值为0,则返回空Tensor。如果self存在元素大于numClasses,这些元素会被编码成全0。
- onValue(aclTensor*,计算输入):公式中的onValue,Device侧的aclTensor,数据类型支持FLOAT16、FLOAT、INT32、INT64,且数据类型需与out一致,支持非连续的Tensor,数据格式支持ND。
- offValue(aclTensor*,计算输入):公式中的offValue,Device侧的aclTensor,数据类型支持FLOAT16、FLOAT、INT32、INT64,且数据类型需与out一致,支持非连续的Tensor,数据格式支持ND。
- axis(int64_t,计算输入):表示编码向量的插入维度,最小值为-1,最大值为self的维度数。若值为-1,编码向量会往self的最后一维插入。
- out(aclTensor*,计算输出):公式中的输出out,Device侧的aclTensor,数据类型支持FLOAT16、FLOAT、INT32、INT64,shape与在self的shape在axis轴插入numClasses后的shape一致,支持非连续的Tensor,数据格式支持ND。
- workspaceSize(uint64_t*,出参):返回需要在Device侧申请的workspace大小。
- executor(aclOpExecutor**,出参):返回op执行器,包含了算子计算流程。
返回值:
aclnnStatus:返回状态码,具体参见aclnn返回码。
第一段接口完成入参校验,出现如下场景时报错: 返回161001(ACLNN_ERR_PARAM_NULLPTR): 1. 传入的self、onValue、offValue或out为空指针。 返回161002(ACLNN_ERR_PARAM_INVALID): 1. self、onValue、offValue或out不在支持的数据类型范围之内。 2. onValue、offValue的数据类型与out的数据类型不一致。 3. self为空Tensor,且numClasses小于等于0。 4. self不为空Tensor,且numClasses小于0。 5. axis的值小于-1。 6. axis的值大于self的维度数量。 7. out的维度不比self的维度多1维。 8. out的shape与在self的shape在axis轴插入numClasses后的shape不一致。
aclnnOneHot
参数说明:
- workspace(void*,入参):在Device侧申请的workspace内存地址。
- workspaceSize(uint64_t,入参):在Device侧申请的workspace大小,由第一段接口aclnnOneHotGetWorkspaceSize获取。
- executor(aclOpExecutor*,入参):op执行器,包含了算子计算流程。
- stream(aclrtStream,入参):指定执行任务的AscendCL Stream流。
返回值:
aclnnStatus:返回状态码,具体参见aclnn返回码。
约束与限制
无。
调用示例
示例代码如下,仅供参考,具体编译和执行过程请参考编译与运行样例。
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_one_hot.h"
#define CHECK_RET(cond, return_expr) \
do { \
if (!(cond)) { \
return_expr; \
} \
} while (0)
#define LOG_PRINT(message, ...) \
do { \
printf(message, ##__VA_ARGS__); \
} while (0)
int64_t GetShapeSize(const std::vector<int64_t> &shape)
{
int64_t shape_size = 1;
for (auto i : shape) {
shape_size *= i;
}
return shape_size;
}
int Init(int32_t deviceId, aclrtStream* stream) {
// 固定写法,AscendCL初始化
auto ret = aclInit(nullptr);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret);
ret = aclrtSetDevice(deviceId);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret);
ret = aclrtCreateStream(stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret);
return 0;
}
template <typename T>
int CreateAclTensor(const std::vector<T> &hostData, const std::vector<int64_t> &shape, void **deviceAddr,
aclDataType dataType, aclTensor **tensor)
{
auto size = GetShapeSize(shape) * sizeof(T);
// 调用aclrtMalloc申请device侧内存
auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret);
// 调用aclrtMemcpy将host侧数据拷贝到device侧内存上
ret = aclrtMemcpy(*deviceAddr, size, hostData.data(), size, ACL_MEMCPY_HOST_TO_DEVICE);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMemcpy failed. ERROR: %d\n", ret); return ret);
// 计算连续tensor的strides
std::vector<int64_t> strides(shape.size(), 1);
for (int64_t i = shape.size() - 2; i >= 0; i--) {
strides[i] = shape[i + 1] * strides[i + 1];
}
// 调用aclCreateTensor接口创建aclTensor
*tensor = aclCreateTensor(shape.data(),
shape.size(),
dataType,
strides.data(),
0,
aclFormat::ACL_FORMAT_ND,
shape.data(),
shape.size(),
*deviceAddr);
return 0;
}
int main()
{
// 1. (固定写法)device/stream初始化, 参考acl对外接口列表
// 根据自己的实际device填写deviceId
int32_t deviceId = 0;
aclrtStream stream;
auto ret = Init(deviceId, &stream);
// check根据自己的需要处理
CHECK_RET(ret == 0, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);
// 2. 构造输入与输出,需要根据API的接口自定义构造
std::vector<int64_t> selfShape = {4, 2};
int numClasses = 4;
std::vector<int64_t> outShape = {4, 2, 4};
std::vector<int64_t> onValueShape = {1};
std::vector<int64_t> offValueShape = {1};
void *selfDeviceAddr = nullptr;
void *outDeviceAddr = nullptr;
void *onValueDeviceAddr = nullptr;
void *offValueDeviceAddr = nullptr;
aclTensor *self = nullptr;
aclTensor *out = nullptr;
aclTensor *onValue = nullptr;
aclTensor *offValue = nullptr;
std::vector<int32_t> selfHostData = {0, 1, 2, 3, 3, 2, 1, 0};
std::vector<int32_t> outHostData = {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
std::vector<int32_t> onValueHostData = {1};
std::vector<int32_t> offValueHostData = {0};
// 创建self aclTensor
ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_INT32, &self);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建out aclTensor
ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_INT32, &out);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建onValue aclTensor
ret = CreateAclTensor(onValueHostData, onValueShape, &onValueDeviceAddr, aclDataType::ACL_INT32, &onValue);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建offValue aclTensor
ret = CreateAclTensor(offValueHostData, offValueShape, &offValueDeviceAddr, aclDataType::ACL_INT32, &offValue);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 3. 调用CANN算子库API,需要修改为具体的API
uint64_t workspaceSize = 0;
int64_t axis = -1;
aclOpExecutor *executor;
// 调用aclnnoneHot第一段接口
ret = aclnnOneHotGetWorkspaceSize(self, numClasses, onValue, offValue, axis, out, &workspaceSize, &executor);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnOneHotGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
// 根据第一段接口计算出的workspaceSize申请device内存
void *workspaceAddr = nullptr;
if (workspaceSize > 0) {
ret = aclrtMalloc(&workspaceAddr, workspaceSize, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret;);
}
// 调用aclnnOnehot第二段接口
ret = aclnnOneHot(workspaceAddr, workspaceSize, executor, stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnOneHot failed. ERROR: %d\n", ret); return ret);
// 4. (固定写法)同步等待任务执行结束
ret = aclrtSynchronizeStream(stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
// 5. 获取输出的值,将device侧内存上的结果拷贝至host侧,需要根据具体API的接口定义修改
auto size = GetShapeSize(outShape);
std::vector<int32_t> resultData(size, 0);
ret = aclrtMemcpy(resultData.data(),
resultData.size() * sizeof(resultData[0]),
outDeviceAddr,
size * sizeof(int32_t),
ACL_MEMCPY_DEVICE_TO_HOST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
for (int64_t i = 0; i < size; i++) {
LOG_PRINT("result[%ld] is: %d\n", i, resultData[i]);
}
// 6. 释放aclTensor和aclScalar,需要根据具体API的接口定义修改
aclDestroyTensor(self);
aclDestroyTensor(onValue);
aclDestroyTensor(offValue);
aclDestroyTensor(out);
// 7. 释放device资源,需要根据具体API的接口定义修改
aclrtFree(selfDeviceAddr);
aclrtFree(onValueDeviceAddr);
aclrtFree(offValueDeviceAddr);
aclrtFree(outDeviceAddr);
if (workspaceSize > 0) {
aclrtFree(workspaceAddr);
}
aclrtDestroyStream(stream);
aclrtResetDevice(deviceId);
aclFinalize();
return 0;
}