aclnnBinaryCrossEntropyWithLogitsTargetBackward

📄 View Source Code

Applicable Products

Function

Performs logits computation on the input self, and then performs backward propagation computation of BCELoss with respect to target using the computed values and the label values target.

Function Prototype

Each operator consists of a two-phase API. You must first call the "aclnnBinaryCrossEntropyWithLogitsTargetBackwardGetWorkspaceSize" API to obtain the input parameters and calculate the required workspace size based on the process, and then call the "aclnnBinaryCrossEntropyWithLogitsTargetBackward" API to perform the computation.

aclnnStatus aclnnBinaryCrossEntropyWithLogitsTargetBackwardGetWorkspaceSize(
  const aclTensor *gradOutput,
  const aclTensor *self,
  const aclTensor *target,
  const aclTensor *weightOptional,
  const aclTensor *posWeightOptional,
  int64_t          reduction,
  aclTensor       *gradTarget,
  uint64_t        *workspaceSize,
  aclOpExecutor  **executor)

aclnnStatus aclnnBinaryCrossEntropyWithLogitsTargetBackward(
  void          *workspace,
  uint64_t       workspaceSize,
  aclOpExecutor *executor,
  aclrtStream    stream)

aclnnBinaryCrossEntropyWithLogitsTargetBackwardGetWorkspaceSize

• Parameters:

  Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (Shape)	Non-contiguous Tensor
  gradOutput (aclTensor*)	Input	The gradient value from the previous step of network backward propagation.	The shape must be broadcastable to the shape of self.	FLOAT16, FLOAT, BFLOAT16	ND	1-8	√
  self（aclTensor*）	Input	The computation result of the previous layer in the forward network.	-	FLOAT16, FLOAT, BFLOAT16	ND	1-8	√
  target (aclTensor*)	Input	Label value.	The shape must be the same as the shape of self.	FLOAT16, FLOAT, BFLOAT16	ND	Same as self.	√
  weightOptional (aclTensor*)	Input	Binary Cross Entropy weight.	The shape must be broadcast to the shape of self. When weightOptional is empty, a tensor of all ones is created with the shape of self.	FLOAT16, FLOAT, BFLOAT16	ND	1-8	√
  posWeightOptional (aclTensor*)	Input	Weight of the positive class.	The shape can be broadcast to the shape of self. When posWeightOptional is null, a tensor of all ones is created with the shape of self.	FLOAT16, FLOAT, BFLOAT16	ND	1-8	√
  reduction (int64_t)	Input	Specifies the reduce operation performed on the gradient computation result of binary cross entropy backward propagation.	Only supports three values: 0, 1, and 2. 0 indicates no operation; 1 indicates averaging the result; 2 indicates summing the result.	INT64	-	-	-
  gradTarget (aclTensor*)	Output	Stores the gradient computation result.	The dtype derived from the input tensor, which must be convertible to the dtype of gradTarget.	FLOAT16, FLOAT, BFLOAT16	Same as self.	Same as self.	√
  workspaceSize (uint64_t*)	Output	Returns the workspace size to be allocated on the Device side.	-	-	-	-	-
  executor (aclOpExecutor**)	Output	Returns the operator executor, which contains the operator computation flow.	-	-	-	-	-

• Return value

  aclnnStatus: return code. For details, see aclnn Return Code.

  The first-phase API performs input parameter validation and returns an error in the following scenarios:

  Return Value	Error Code	Description
  ACLNN_ERR_PARAM_NULLPTR	161001	The passed gradOutput, self, target, and gradTarget are null pointers.
  ACLNN_ERR_PARAM_INVALID	161002	The data types and data formats of gradOutput, self, target, and gradTarget are not within the supported range.
  		When weightOptional and posWeightOptional are not null pointers, their data type and data format are not within the supported range.
  		The shapes of self, target, and gradTarget are inconsistent.
  		When weightOptional and posWeightOptional are not null pointers, their shapes cannot be broadcast to the shape of self.
  		gradOutput cannot be broadcast to the shape of self.
  		The value of reduction is not one of 0, 1, or 2.
  		The dimensions of gradOutput, self, target, and gradTarget are greater than 8.

aclnnBinaryCrossEntropyWithLogitsTargetBackward

• Parameters:

  Parameter	Input/Output	Description
  workspace	Input	Specifies the workspace memory address applied for on the device.
  workspaceSize	Input	Specifies the workspace size applied for on the device, which is obtained through the first-phase API aclnnBinaryCrossEntropyWithLogitsTargetBackwardGetWorkspaceSize.
  executor	Input	Operator executor, which contains the operator computation process.
  stream	Input	Specifies the stream that executes the task.

• Return value

  aclnnStatus: return code. For details, see aclnn Return Code.

Constraints

• Deterministic computation:

  • aclnnBinaryCrossEntropyWithLogitsTargetBackward defaults to a deterministic implementation.

Example

The following provides example code for reference only. For details about compilation and running, see Compile and Run Samples.

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_binary_cross_entropy_with_logits_target_backward.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 shapeSize = 1;
  for (auto i : shape) {
    shapeSize *= i;
  }
  return shapeSize;
}

int Init(int32_t deviceId, aclrtStream* stream) {
  // Boilerplate for resource initialization.
  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);
  // Call aclrtMalloc to apply for device-side memory.
  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);

  // Copy data from the host to the device using aclrtMemcpy.
  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);

  // Compute strides for a contiguous tensor.
  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];
  }

  // Create an aclTensor using the aclCreateTensor API.
  *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
                            shape.data(), shape.size(), *deviceAddr);
  return 0;
}

int main() {
  // 1. (Boilerplate) Initialize the device/stream. Refer to the ACL API manual.
  // Specify the deviceId based on your actual device.
  int32_t deviceId = 0;
  aclrtStream stream;
  auto ret = Init(deviceId, &stream);
  // Handle the check based on your needs.
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);

  // 2. Construct the inputs and outputs. Customize the construction based on the API.
  std::vector<int64_t> gradOutputShape = {2, 2};
  std::vector<int64_t> selfShape = {2, 2};
  std::vector<int64_t> targetShape = {2, 2};
  std::vector<int64_t> weightShape = {2, 2};
  std::vector<int64_t> posWeightShape = {2, 2};
  std::vector<int64_t> gradTargetShape = {2, 2};
  void* gradOutputDeviceAddr = nullptr;
  void* selfDeviceAddr = nullptr;
  void* targetDeviceAddr = nullptr;
  void* weightDeviceAddr = nullptr;
  void* posWeightDeviceAddr = nullptr;
  void* gradTargetDeviceAddr = nullptr;
  aclTensor* gradOutput = nullptr;
  aclTensor* self = nullptr;
  aclTensor* target = nullptr;
  aclTensor* gradTarget = nullptr;
  aclTensor* weight = nullptr;
  aclTensor* posWeight = nullptr;
  std::vector<float> gradOutputHostData = {0, 1, 2, 3};
  std::vector<float> selfHostData = {0, 1, 2, 3};
  std::vector<float> targetHostData = {0.1, 0.1, 0.1, 0.1};
  std::vector<float> weightHostData = {0, 1, 2, 3};
  std::vector<float> posWeightHostData = {0, 1, 2, 3};
  std::vector<float> gradTargetHostData = {0, 0, 0, 0};
  int64_t reduction = 0;

  // Create the gradOutput aclTensor.
  ret = CreateAclTensor(gradOutputHostData, gradOutputShape, &gradOutputDeviceAddr, aclDataType::ACL_FLOAT, &gradOutput);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the self aclTensor.
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the target aclTensor.
  ret = CreateAclTensor(targetHostData, targetShape, &targetDeviceAddr, aclDataType::ACL_FLOAT, &target);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the weight aclTensor.
  ret = CreateAclTensor(weightHostData, weightShape, &weightDeviceAddr, aclDataType::ACL_FLOAT, &weight);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the posWeight aclTensor.
  ret = CreateAclTensor(posWeightHostData, posWeightShape, &posWeightDeviceAddr, aclDataType::ACL_FLOAT, &posWeight);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the gradTarget aclTensor.
  ret = CreateAclTensor(gradTargetHostData, gradTargetShape, &gradTargetDeviceAddr, aclDataType::ACL_FLOAT, &gradTarget);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;

  // Example of aclnnBinaryCrossEntropyWithLogitsTargetBackward API call
  // 3. Call the CANN operator library API. Replace with the specific API name.
  // Call the first-phase API of aclnnBinaryCrossEntropyWithLogitsTargetBackward.
  ret = aclnnBinaryCrossEntropyWithLogitsTargetBackwardGetWorkspaceSize(gradOutput, self, target, weight, posWeight,
      reduction, gradTarget, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnBinaryCrossEntropyWithLogitsTargetBackwardGetWorkspaceSize failed. ERROR: %d\n",
                                          ret); return ret);
  // Allocate device memory based on the workspaceSize calculated by the first-phase API.
  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);
  }
  // Call the second-phase API of aclnnBinaryCrossEntropyWithLogitsTargetBackward.
  ret = aclnnBinaryCrossEntropyWithLogitsTargetBackward(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnBinaryCrossEntropyWithLogitsTargetBackward failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate) Synchronously wait for task execution to complete.
  ret = aclrtSynchronizeStream(stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);

  // 5. Obtain the output value and copy the result from the Device side memory to the Host side. Modify this based on the specific API definition.
  auto size = GetShapeSize(gradTargetShape);
  std::vector<float> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), gradTargetDeviceAddr,
                    size * sizeof(resultData[0]), 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. Release aclTensor and aclScalar. Modify based on the specific API definition.
  aclDestroyTensor(gradOutput);
  aclDestroyTensor(self);
  aclDestroyTensor(target);
  aclDestroyTensor(weight);
  aclDestroyTensor(posWeight);
  aclDestroyTensor(gradTarget);

  // 7. Release device resources. Modify based on the specific API definition.
  aclrtFree(gradOutputDeviceAddr);
  aclrtFree(selfDeviceAddr);
  aclrtFree(targetDeviceAddr);
  aclrtFree(weightDeviceAddr);
  aclrtFree(posWeightDeviceAddr);
  aclrtFree(gradTargetDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
  return 0;
}