ReductionConfig
API Definition |
inline ReductionConfig(std::vector<int> deviceList, const char *model, uint64_t modelSize) |
|---|---|
Function |
Default constructor of ReductionConfig. This function is used when NN is used to reduce dimensions. |
Input |
std::vector<int> deviceList: resource configuration on the NPU. const char *model: neural network dimension reduction model. uint64_t modelSize: model size. |
Output |
N/A |
Return Value |
N/A |
Restrictions |
When different dimension reduction algorithms are used, corresponding parameters need to be configured, and the dimensions after dimension reduction must meet the dimension restrictions of the dimension-reduced data index. When NN is used, ensure that the model path exists, the model is correctly used, the model length is equal to modelSize, and the model size is less than or equal to 128 MB. |