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 of backpropagation, aclTensor on the device. The data type can be FLOAT16, FLOAT32, DOUBLE, COMPLEX64, or COMPLEX128. The shape can be 4D or 5D, must be the same as that of[object Object]and[object Object], and also must be the same as that of output of[object Object]forward propagation. The can be ND.[object Object](aclTensor*, computation input): input tensor of forward propagation, aclTensor on the device. The data type can be FLOAT16, FLOAT32, DOUBLE, COMPLEX64, or COMPLEX128. The shape can be 4D or 5D, with the same dimensions as[object Object]and[object Object]. The shape must be the same as that of[object Object]. The can be ND.[object Object](aclIntArray*, computation input): aclIntArray on the device. The length is 6, with values to be padded on the left, right, top, and bottom sides in sequence. The first two values of[object Object]must be less than the value of the last dimension of[object Object], the middle two values must be less than the value of the second-to-last dimension of[object Object], and the last two values must be less than the value of the third-to-last dimension of[object Object].[object Object](aclTensor*, computation output): output tensor of backward propagation, aclTensor on the device. The data type can be FLOAT16, FLOAT32, DOUBLE, COMPLEX64, or COMPLEX128. The shape can be 4D or 5D, with the same dimensions as[object Object]and[object Object]. The can be ND.[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]
Parameters:
[object Object](void*, input): address of the workspace to be allocated on the device.[object Object](uint64_t, input): size of the workspace to be allocated on the device, which is obtained by calling the first-phase API[object Object].[object Object](aclOpExecutor*, input): operator executor, containing the operator computation process.[object Object](aclrtStream, input): stream for executing the task.
Returns:
- Deterministic computation:
[object Object]defaults to deterministic implementation. If the number of elements in[object Object]is greater than 300*1024*1024, the execution may time out.
The following example is for reference only. For details, see .
[object Object]