Description: Performs backward propagation of reflection_pad2d. For details about the forward propagation, see .
Example:
[object Object]
Each operator has calls. First, [object Object] is called to obtain the workspace size required for computation and the executor that contains the operator computation process. Then, [object Object] is called to perform computation.
[object Object][object Object]
Parameters:
[object Object](aclTensor*, computation input): input gradient, aclTensor on the device. The data type can be BFLOAT16, FLOAT16, FLOAT, DOUBLE, COMPLEX64, or COMPLEX128, which is the same as that of[object Object]. are supported. The can be ND. The shape supports 3–4 dimensions, which must be the same as those of[object Object]and[object Object]. The shape must be the same as that of the[object Object]in forward propagation[object Object].[object Object](aclTensor*, computation input): tensor to be padded during forward propagation. The data type can be BFLOAT16, FLOAT16, FLOAT, DOUBLE, COMPLEX64, or COMPLEX128. It is an aclTensor on the device. are supported. The can be ND. The shape supports 3–4 dimensions, which must be the same as those of[object Object]and[object Object]. The shape is the same as that of[object Object].[object Object](aclIntArray*, computation input): padding range, aclIntArray on the device. The data type is INT64 and the length is 4. The values of the first two dimensions of[object Object]must be less than the value of the last dimension of[object Object], and the values of the last two dimensions of[object Object]must be less than the value of the penultimate dimension of[object Object].[object Object](aclTensor*, computation output): gradient of[object Object], aclTensor on the device. The data type can be BFLOAT16, FLOAT16, FLOAT, DOUBLE, COMPLEX64, or COMPLEX128, which is the same as that of[object Object]. The shape is the same as that of[object Object]. are supported. The can be ND. The shape supports 3–4 dimensions, which must be the same as those of[object Object]and[object Object]. The shape is the same as that of[object Object].[object Object](uint64_t*, output): size of the workspace to be allocated on the device.[object Object](aclOpExecutor**, output): operator executor, containing the operator computation process.
Returns:
[object Object]