AscendNNInference

AscendNNInference(std::vector<int> deviceList, const char* model, uint64_t modelSize);

Constructor of AscendNNInference, which generates AscendNNInference. The Ascend AI Processor resources on the device and model path are set based on the values configured in deviceList.

std::vector<int> deviceList: device ID

const char* model: neural dimension reduction model.

uint64_t modelSize: size of the neural dimension reduction model.

N/A

N/A

• Value range of deviceList: (0, 32].
• model must be a valid memory pointer of the neural dimension reduction model. modelSize indicates its size and the value range is (0, 128 MB]. If the parameters do not match, model instantiation or inference may fail. An invalid model may damage the system. Ensure that the model source is valid.
  • dimsIn ∈ {64, 128, 256, 384, 512, 768, 1024}
  • dimsOut ∈ {32, 64, 96, 128, 256}
  • batches ∈ {1, 2, 4, 8, 16, 32, 64, 128}

AscendNNInference(const AscendNNInference&) = delete;

Declares that the copy constructor of this Index class is void, that is, it cannot be copied.

const AscendNNInference&: constant AscendNNInference

N/A

N/A

N/A