NVIDIA Holoscan SDK v0.5.1
Holoscan v0.5.1

Struct MultiAISpecs

struct MultiAISpecs

Struct that holds specifications related to Multi AI inference, along with input and output data buffer.

Public Functions

MultiAISpecs() = default

inline MultiAISpecs(const std::string &backend, const Mappings &model_path_map, const Mappings &inference_map, bool is_engine_path, bool oncpu, bool parallel_proc, bool use_fp16, bool cuda_buffer_in, bool cuda_buffer_out)

Constructor.

Parameters
  • backend – Backend inference (trt or onnxrt)

  • model_path_map – Map with model name as key, path to model as value

  • inference_map – Map with model name as key, output tensor name as value

  • oncpu – Perform inference on CPU

  • parallel_proc – Perform parallel inference of multiple models

  • use_fp16 – Use FP16 conversion, only supported for trt

  • cuda_buffer_in – Input buffers on CUDA

  • cuda_buffer_out – Output buffers on CUDA

inline Mappings get_path_map() const

Get the model data path map.

Returns

Mappings data

Public Members

std::string backend_type_ = {"trt"}

Mappings model_path_map_

Map with key as model name and value as model file path.

Mappings inference_map_

Map with key as model name and value as inferred tensor name.

bool is_engine_path_ = false

Flag showing if input model path is path to engine files.

bool oncuda_ = true

Flag showing if inference on CUDA. Default is True.

bool parallel_processing_ = false

Flag to enable parallel inference. Default is True.

bool use_fp16_ = false

Flag showing if trt engine file conversion will use FP16. Default is False.

bool cuda_buffer_in_ = true

Flag showing if input buffers are on CUDA. Default is True.

bool cuda_buffer_out_ = true

Flag showing if output buffers are on CUDA. Default is True.

DataMap data_per_model_

Input Data Map with key as model name and value as DataBuffer.

DataMap data_per_tensor_

Input Data Map with key as tensor name and value as DataBuffer.

DataMap output_per_model_

Output Data Map with key as tensor name and value as DataBuffer.

© Copyright 2022-2023, NVIDIA. Last updated on Jul 28, 2023.