aclnnDropoutBackward

📄 View Source Code

Applicable Products

Function

• Description: During training, based on the value of the corresponding bit in the mask, elements in the input are set to zero and scaled up or zeroed according to scale. If the corresponding bit in the mask is 1, the corresponding element in gradOutput is scaled up. If the bit in the mask is 0, the corresponding element in gradOutput is set to zero. In particular, if scale is 0, all elements are set to 0; if scale is 1, the elements of gradOutput remain unchanged. It is recommended to use this API in combination with aclnnDropoutGenMask or aclnnDropoutGenMaskV2.

• Formula:

  $$out_i=\begin{cases} 0,&\text { with probability }1-1/scale \\ scale * self_i, &\text { with probability }1/scale \end{cases}$$

Function Prototype

Each operator uses a two-phase API. You must first call the "aclnnDropoutBackwardGetWorkspaceSize" API to obtain the workspace size required for computation and the executor that encapsulates the operator's computation process, and then call the "aclnnDropoutBackward" API to perform the computation.

aclnnStatus aclnnDropoutBackwardGetWorkspaceSize(
  const aclTensor* gradOutput,
  const aclTensor* mask,
  double           scale,
  aclTensor*       out,
  uint64_t*        workspaceSize,
  aclOpExecutor**  executor)

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

aclnnDropoutBackwardGetWorkspaceSize

• Parameters:

  Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (shape)	Non-contiguous Tensor
  gradOutput (aclTensor*)	Input	Represents the tensor to be processed.	Does not support empty tensors.	FLOAT, FLOAT16, BFLOAT16	ND	1-8	√
  mask（aclTensor*）	Input	Mask data of bit type stored in UINT8 type.	Empty tensors are not supported. The UINT8 data type is supported. The shape must be (align(number of elements in gradOutput, 128)/8), indicating that the number of bits must be aligned to 128. Here, align means rounding up the number of elements in input to the nearest multiple of 128. For example, align(1111, 128) yields 1152.	UINT8	ND	1-8	√
  scale (double)	Input	Input scale, which specifies the scaling ratio for the output data.	The data type supports DOUBLE. scale must be 0 or greater than or equal to 1.	-	-	-	-
  out (aclTensor*)	Output	Output out.	Empty tensors are not supported. The data type must be convertible from gradOutput The shape must be the same as gradOutput	FLOAT, FLOAT16, BFLOAT16	ND	1-8	√
  workspaceSize (uint64_t*)	Output	Returns the workspace size to be requested on the Device side.	-	-	-	-	-
  executor (aclOpExecutor**)	Output	Returns the operator executor, which contains the operator computation process.	-	-	-	-	-

  • Atlas training products: The BFLOAT16 data type is not supported.

• Return value

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

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

  Return Code	Error Code	Description
  ACLNN_ERR_PARAM_NULLPTR	161001	Indicates that gradOutput, mask, or out is a null pointer.
  ACLNN_ERR_PARAM_INVALID	161002	The data types of gradOutput, mask, and out are not within the supported range.
  		scale is less than 1 and not equal to 0.
  		The dimension of gradOutput exceeds 8.
  		The shapes of gradOutput and out are inconsistent.
  		The shape of mask does not meet the condition.

aclnnDropoutBackward

• Parameters:

  Parameter	Input/Output	Description
  workspace	Input	Specifies the workspace memory address applied for on the Device side.
  workspaceSize	Input	Specifies the workspace size applied for on the Device side, which is obtained by the first-phase API aclnnDropoutBackwardGetWorkspaceSize.
  executor	Input	Operator executor that contains the operator computation flow.
  stream	Input	Specifies the stream for task execution.

• Return value

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

Constraints

• Deterministic computation:

  • aclnnDropoutBackward 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_dropout_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: 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 allocate 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);
  // Call aclrtMemcpy to copy data from host to device memory.
  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 the strides of 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];
  }

  // Call the aclCreateTensor API to create an 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. (Boilerplate) Initialize the device/stream. See the ACL API manual.
  // Enter the deviceId based on the actual device.
  int32_t deviceId = 0;
  aclrtStream stream;
  auto ret = Init(deviceId, &stream);
  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 = {4, 2};
  std::vector<int64_t> maskShape = {16};
  std::vector<int64_t> outShape = {4, 2};
  void* gradOutputDeviceAddr = nullptr;
  void* maskDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* gradOutput = nullptr;
  aclTensor* mask = nullptr;
  aclTensor* out = nullptr;
  std::vector<float> gradOutputHostData = {0, 1, 2, 3, 4, 5, 6, 7};
  std::vector<uint8_t> maskHostData(16, 0);
  maskHostData[0] = 42;
  std::vector<float> outHostData(8, 0);
  double scale = 2;
  // Create the gradOutput aclTensor.
  ret =
      CreateAclTensor(gradOutputHostData, gradOutputShape, &gradOutputDeviceAddr, aclDataType::ACL_FLOAT, &gradOutput);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the mask aclTensor.
  ret = CreateAclTensor(maskHostData, maskShape, &maskDeviceAddr, aclDataType::ACL_UINT8, &mask);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the out aclTensor.
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  // 3. Call the CANN operator library API. Replace with the actual API name.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnDropoutBackward.
  ret = aclnnDropoutBackwardGetWorkspaceSize(gradOutput, mask, scale, out, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnDropoutBackwardGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // Apply for 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 aclnnDropoutBackward.
  ret = aclnnDropoutBackward(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnDropoutBackward 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 memory to the host. Modify based on the specific API definition.
  auto size = GetShapeSize(outShape);
  std::vector<float> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
                    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(mask);
  aclDestroyTensor(out);

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