aclnnNLLLoss2d
接口原型
每个算子有两段接口，必须先调用“aclnnXxxGetWorkspaceSize”接口获取入参并根据计算流程计算所需workspace大小，再调用“aclnnXxx”接口执行计算。两段式接口如下：
- 第一段接口：aclnnStatus aclnnNLLLoss2dGetWorkspaceSize(const aclTensor *self, const aclTensor *target, const aclTensor *weight, int64_t reduction, int64_t ignoreIndex, aclTensor *out, aclTensor *totalWeightOut, uint64_t *workspaceSize, aclOpExecutor **executor)
- 第二段接口：aclnnStatus aclnnNLLLoss2d(void* workspace, uint64_t workspaceSize, aclOpExecutor* executor, aclrtStream stream)
功能描述
aclnnNLLLoss2dGetWorkspaceSize
- 接口定义：
aclnnStatus aclnnNLLLoss2dGetWorkspaceSize(const aclTensor *self, const aclTensor *target, const aclTensor *weight, int64_t reduction, int64_t ignoreIndex, aclTensor *out, aclTensor *totalWeightOut, uint64_t *workspaceSize, aclOpExecutor **executor)
- 参数说明：
- self：Device侧的aclTensor，shape为4维，第1、2维是N和C，分别表示batch size和类别数。数据类型支持FLOAT。支持非连续的Tensor，数据格式支持ND。
- target：Device侧的aclTensor，表示真实标签，shape为3维，第1维是N，其中每个元素的取值范围是[0, C-1]。数据类型支持INT64、UINT8。支持非连续的Tensor，数据格式支持ND。
- weight：Device侧的aclTensor，表示每个类别的缩放权重，shape为(C, )，数据类型支持FLOAT。支持非连续的Tensor，数据格式支持ND。
- reduction：Host侧的int64_t，指定要应用到输出的缩减。支持3种枚举值：当取值为0，即为"none"，表示不应用减少；当取值为1，即为"mean"，表示输出的总和将除以输出中的元素数；当取值为2，即为"sum"， 表示输出将被求和。
- ignoreIndex：Host侧的int64_t，指定一个被忽略且不影响输入梯度的目标值。
- out：Device侧的aclTensor，数据类型支持FLOAT且与self相同，shape(N, )为或者(1, )，数据格式支持ND，且数据格式需要与self一致。
- totalWeightOut：Device侧的aclTensor，数据类型支持FLOAT且与weight相同，shape为(1, )，数据格式支持ND。
- workspaceSize：返回用户需要在Device侧申请的workspace大小。
- executor：返回op执行器，包含了算子计算流程。
- 返回值：
返回aclnnStatus状态码，具体参见aclnn返回码。
第一段接口完成入参校验，出现以下场景时报错：
- 返回161001（ACLNN_ERR_PARAM_NULLPTR）：传入的self、target、weight、out或totalWeightOut是空指针。
- 返回161002（ACLNN_ERR_PARAM_INVALID）：
- self、target、weight、out或totalWeightOut的数据类型和数据格式不在支持的范围之内。
- self、weight、out或totalWeightOut的数据类型不一致。
- self、target、weight、out、totalWeightOut的shape和format不正确。
- reduction值不在0~2范围之内。
aclnnNLLLoss2d
- 接口定义：
aclnnStatus aclnnNLLLoss2d(void* workspace, uint64_t workspaceSize, aclOpExecutor* executor, aclrtStream stream)
- 参数说明：
- workspace：在Device侧申请的workspace内存起址。
- workspaceSize：在Device侧申请的workspace大小，由第一段接口aclnnNLLLoss2dGetWorkspaceSize获取。
- executor：op执行器，包含了算子计算流程。
- stream：指定执行任务的AscendCL stream流。
- 返回值：
返回aclnnStatus状态码，具体参见aclnn返回码。
调用示例
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_nll_loss2d.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, aclrtContext* context, 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 = aclrtCreateContext(context, deviceId);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateContext failed. ERROR: %d\n", ret); return ret);
ret = aclrtSetCurrentContext(*context);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetCurrentContext 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/context/stream初始化, 参考AscendCL对外接口列表
// 根据自己的实际device填写deviceId
int32_t deviceId = 0;
aclrtContext context;
aclrtStream stream;
auto ret = Init(deviceId, &context, &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 = {1, 2, 3, 2};
std::vector<int64_t> targetShape = {1, 3, 2};
std::vector<int64_t> weightShape = {2};
std::vector<int64_t> outShape = {1, 3, 2};
std::vector<int64_t> totalWeightOutShape = {1};
void* selfDeviceAddr = nullptr;
void* targetDeviceAddr = nullptr;
void* weightDeviceAddr = nullptr;
void* outDeviceAddr = nullptr;
void* totalWeightOutDeviceAddr = nullptr;
aclTensor* self = nullptr;
aclTensor* target = nullptr;
aclTensor* weight = nullptr;
aclTensor* out = nullptr;
aclTensor* totalWeightOut = nullptr;
std::vector<float> selfHostData = {0.1, 1.1, 2.1, 3.1, 4.1, 5.1, 6.1, 7.1, 8.1, 9.1, 10.1, 11.1};
std::vector<int32_t> targetHostData = {1, 0, 1, 1, 2, 1};
std::vector<float> weightHostData = {1.1, 1.2};
std::vector<float> outHostData = {0.0, 0.0, 0.0, 0.0, 0.0, 0.0};
std::vector<float> totalWeightOutHostData = {0};
int64_t reduction = 0;
int64_t ignoreIndex = -100;
// 创建self aclTensor
ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建target aclTensor
ret = CreateAclTensor(targetHostData, targetShape, &targetDeviceAddr, aclDataType::ACL_INT32, &target);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建weight aclScalar
ret = CreateAclTensor(weightHostData, weightShape, &weightDeviceAddr, aclDataType::ACL_FLOAT, &weight);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建out aclTensor
ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建totalWeightOut aclTensor
ret = CreateAclTensor(totalWeightOutHostData, totalWeightOutShape, &totalWeightOutDeviceAddr, aclDataType::ACL_FLOAT,
&totalWeightOut);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 3.调用CANN算子库API，需要修改为具体的算子接口
uint64_t workspaceSize = 0;
aclOpExecutor* executor;
// 调用aclnnNLLLoss2d第一段接口
ret = aclnnNLLLoss2dGetWorkspaceSize(self, target, weight, reduction, ignoreIndex, out, totalWeightOut, &workspaceSize,
&executor);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnNLLLoss2dGetWorkspaceSize 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aclnnNLLLoss2d第二段接口
ret = aclnnNLLLoss2d(workspaceAddr, workspaceSize, executor, stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnNLLLoss2d 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<float> resultData(size, 0);
ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(float),
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: %f\n", i, resultData[i]);
}
// 6. 释放aclTensor，需要根据具体API的接口定义修改
aclDestroyTensor(self);
aclDestroyTensor(target);
aclDestroyTensor(weight);
aclDestroyTensor(out);
return 0;
}
父主题： NN类算子接口