aclnnUpsampleBicubic2dBackward

Supported Products

Product	Supported
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	×
Atlas inference products	×
Atlas training products	√

Function

API description: Performs backpropagation of aclnnUpsampleBicubic2d. If the shape of the input tensor is (N, C, H, W), then the shape of the output tensor is (N, C, inputSize[2], inputSize[3]).
Formula: For a two-dimensional interpolation point $(N, C, h, w)$ , the interpolation $gradInput(N, C, h, w)$ may be represented as:
${gradInput(N, C, h, w)}=\sum_{i=0}^{3}\sum_{j=0}^{3}{W(i, j)}*{f(h_i, w_j)}$ $scaleH =\begin{cases} (inputSize[2]-1) / (outputSize[0]-1) & alignCorners=true \\ 1 / scalesH & alignCorners=false\&scalesH>0\\ inputSize[2] / outputSize[0] & otherwise \end{cases}$ $scaleW =\begin{cases} (inputSize[3]-1) / (outputSize[1]-1) & alignCorners=true \\ 1 / scalesW & alignCorners=false\&scalesW>0\\ inputSize[3] / outputSize[1] & otherwise \end{cases}$
Where,
- i and j are index variables of $W(i, j)$ .
- $f(h_i, w_j)$ is the pixel value of gradOut in $(h_i, w_j)$ .
- $W(i, j)$ is the weight of the bicubic anti-aliasing interpolation, which is defined as follows: $W(d) =\begin{cases} (a+2)|d|^3-(a+3)|d|^2+1 & |d|\leq1 \\ a|d|^3-5a|d|^2+8a|d|-4a & 1<|d|<2 \\ 0 & otherwise \end{cases}$ Where,
  - $a=-0.75$
  - $d = |(h, w) - (h_i, w_j)|$

Prototype

Each operator has two-phase API calls. First, aclnnUpsampleBicubic2dBackwardGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnUpsampleBicubic2dBackward is called to perform computation.

aclnnStatus aclnnUpsampleBicubic2dBackwardGetWorkspaceSize(
  const aclTensor*   gradOut,
  const aclIntArray* outputSize,
  const aclIntArray* inputSize,
  const bool         alignCorners,
  double             scalesH,
  double             scalesW,
  aclTensor*         gradInput,
  uint64_t*          workspaceSize,
  aclOpExecutor**    executor)

aclnnStatus aclnnUpsampleBicubic2dBackward(
  void          *workspace,
  uint64_t       workspaceSize,
  aclOpExecutor *executor,
  aclrtStream    stream)

aclnnUpsampleBicubic2dBackwardGetWorkspaceSize

Parameters:

Name	Input/Output	Description	Precaution	Data Type	Data Format	Dimension (Shape)	Non-contiguous Tensor
gradOut	Input	Gradient tensor for backpropagation, corresponding to `gradOut` in the formula description.	Empty tensors are not supported. The data type must be the same as that of `gradInput`. The ND format is processed as NCHW format by default. The values of gradOutput in all dimensions must be less than or equal to (2^31 – 1).	FLOAT32, FLOAT16, BFLOAT16	NCHW, ND, NHWC	4	√
outputSize	Input	Spatial size of the input `gradOut` in the H and W dimensions. It corresponds to `outputSize` in the formula.	The size is 2, and each element is greater than 0.	INT64	-	-	-
inputSize	Input	Spatial size of the output `gradInput` in the N, C, H, and W dimensions or in the N, H, W, and C dimensions. It corresponds to `inputSize` in the formula.	The size is 4, and each element is greater than 0.	INT64	-	-	-
alignCorners	Input	Whether to align corner pixels, corresponding to `alignCorners` in the formula.	If the value is True, the corner pixels of the input and output tensors are aligned. Otherwise, the corner pixels are not aligned.	BOOL	-	-	-
scalesH	Input	Multiplier for the output `gradInput` in the height dimension, corresponding to `scalesH` in the formula.	-	DOUBLE	-	-	-
scalesW	Input	Multiplier for the output `gradInput` in the width dimension, corresponding to `scalesW` in the formula.	-	DOUBLE	-	-	-
gradInput	Output	Output tensor for backpropagation, corresponding to `gradInput` in the formula.	Empty tensors are not supported. The data type must be the same as that of `gradOut`. The values of out in all dimensions must be less than or equal to (2^31 – 1).	FLOAT32, FLOAT16, BFLOAT16	NCHW, ND, NHWC	4	√
workspaceSize	Output	Size of the workspace required to be allocated on the device.	-	-	-	-	-
executor	Output	Operator executor, containing the operator computation process.	-	-	-	-	-

Atlas training products:
- The data types of gradOut and gradInput do not support BFLOAT16.
- The data formats of gradOut and gradInput do not support NHWC.
Atlas A2 training products/Atlas A2 inference products and Atlas A3 training products/Atlas A3 inference products:

The data formats of gradOut and gradInput do not support NHWC.

Returns:

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

The first-phase API implements input parameter verification. The following errors may be thrown.

Return	Error Code	Description
ACLNN_ERR_PARAM_NULLPTR	161001	The passed gradOut, outputSize, inputSize, or gradInput is a null pointer.
ACLNN_ERR_PARAM_INVALID	161002	The data type or format of gradOut is not supported.
		The data types of gradOut and gradInput are inconsistent.
		The shape of gradOut is not 4D.
		The size of outputSize is not 2.
		The value of an element of outputSize is less than 1.
		The size of inputSize is not 4.
		The value of an element of inputSize is less than 1.
		The sizes of gradOut and inputSize in the N and C dimensions are inconsistent.
		The sizes of gradOut in the H and W dimensions are inconsistent with those of outputSize[0] and outputSize[1].
		The dimension sizes of the N or C axes of gradOut and gradInput are different.
		The data format of gradOut or gradInput is not supported.

aclnnUpsampleBicubic2dBackward

Parameters:

Name	Input/Output	Description
workspace	Input	Address of the workspace to be allocated on the device.
workspaceSize	Input	Size of the workspace to be allocated on the device, which is obtained by calling the first-phase API aclnnUpsampleBicubic2dBackwardGetWorkspaceSize.
executor	Input	Operator executor, containing the operator computation process.
stream	Input	Stream for executing the task.

Returns:

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

Constraints

Either the H and W axes of the outputSize parameter or the scalesH and scalesW parameters can be used.
- When alignCorners is set to True:
  - If the value of the corresponding axis of outputSize is equal to 1, the value of the corresponding axis of scales is 0.
  - In other cases, the values of the corresponding axes in the input parameters inputSize and outputSize are used, and $scales = (inputSize – 1)/(outputSize – 1)$ .
- If alignCorners is set to False:
  - If the value of scalesH or scalesW is less than or equal to 0, the value of the corresponding axis in outputSize is used, that is, $scales = (inputSize/outputSize)$ .
  - If the value of scalesH or scalesW is greater than 0, the value of scalesH or scalesW is used. That is, the value of the corresponding axis of outputSize is $floor(inputSize\_H * scalesH)$ or $floor(inputSize\_W * scalesW)$ .
Deterministic computing:

aclnnUpsampleBicubic2dBackward defaults to a non-deterministic implementation. You can call aclrtCtxSetSysParamOpt to enable deterministic computing. A deterministic implementation must meet the following conditions:
- inputSize[3] > 130000
- scaleH >=50
- scaleW >=50 && inputSize[0] * inputSize[1] * inputSize[2] > inputSize[3] * 0.5
- scaleH < 0.02 && scaleW < 0.02 && inputSize[0] * inputSize[1] * inputSize[2] > inputSize[3] * 10000

Example

The following example is for reference only. For details, see Compilation and Running Sample.

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_upsample_bicubic_2d_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)
{
    // (Fixed writing) Initialize resources.
    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 memory on the device.
    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 the data on the host to the memory on the device.
    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 the 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 aclCreateTensor to create an aclTensor.
    *tensor = aclCreateTensor(shape.data(),
        shape.size(),
        dataType,
        strides.data(),
        0,
        aclFormat::ACL_FORMAT_NCHW,
        shape.data(),
        shape.size(),
        *deviceAddr);
    return 0;
}

int main()
{
    // 1. (Fixed writing) Initialize the device and stream. For details, see the ACL API manual.
    // Set the device ID in use.
    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 input and output based on the API.
    std::vector<int64_t> selfShape = {1, 1, 2, 2};
    std::vector<int64_t> outShape = {1, 1, 3, 3};
    void *selfDeviceAddr = nullptr;
    void *outDeviceAddr = nullptr;
    aclTensor *self = nullptr;
    aclTensor *out = nullptr;
    std::vector<float> selfHostData = {1, 2, 3, 4.1};
    std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0, 0};

    // Create a self aclTensor.
    ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // Create an out aclTensor.
    ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    std::vector<int64_t> outArraySize = {2, 2};
    const aclIntArray *outputSize = aclCreateIntArray(outArraySize.data(), outArraySize.size());
    CHECK_RET(outputSize != nullptr, return ACL_ERROR_INTERNAL_ERROR);

    std::vector<int64_t> inputArraySize = {1, 1, 3, 3};
    const aclIntArray *inputSize = aclCreateIntArray(inputArraySize.data(), inputArraySize.size());
    CHECK_RET(inputSize != nullptr, return ACL_ERROR_INTERNAL_ERROR);

    // 3. Call the CANN operator library API, which needs to be replaced with the actual API.
    uint64_t workspaceSize = 0;
    aclOpExecutor *executor;
    // Call the first-phase API of aclnnUpsampleBicubic2dBackward.
    ret = aclnnUpsampleBicubic2dBackwardGetWorkspaceSize(
        self, outputSize, inputSize, 1, 1.1, 1.1, out, &workspaceSize, &executor);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnUpsampleBicubic2dBackwardGetWorkspaceSize failed. ERROR: %d\n", ret);
              return ret);
    // Allocate device memory based on the computed workspaceSize.
    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 aclnnUpsampleBicubic2dBackward.
    ret = aclnnUpsampleBicubic2dBackward(workspaceAddr, workspaceSize, executor, stream);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnUpsampleBicubic2dBackward failed. ERROR: %d\n", ret); return ret);

    // 4. (Fixed writing) Wait until the task execution is 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 to the host. Modify the code based on the 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. Modify the configuration based on the API definition.
    aclDestroyTensor(self);
    aclDestroyTensor(out);
    aclDestroyIntArray(outputSize);
    aclDestroyIntArray(inputSize);

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