Class anira::LibtorchProcessor#

class LibtorchProcessor : public anira::BackendBase#

Inheritance diagram for anira::LibtorchProcessor:

digraph { graph [bgcolor="#00000000"] node [shape=rectangle style=filled fillcolor="#FFFFFF" font=Helvetica padding=2] edge [color="#1414CE"] "2" [label="anira::BackendBase" tooltip="anira::BackendBase"] "1" [label="anira::LibtorchProcessor" tooltip="anira::LibtorchProcessor" fillcolor="#BFBFBF"] "1" -> "2" [dir=forward tooltip="public-inheritance"] }

Collaboration diagram for anira::LibtorchProcessor:

digraph { graph [bgcolor="#00000000"] node [shape=rectangle style=filled fillcolor="#FFFFFF" font=Helvetica padding=2] edge [color="#1414CE"] "8" [label="anira::MemoryBlock< float >" tooltip="anira::MemoryBlock< float >"] "2" [label="anira::BackendBase" tooltip="anira::BackendBase"] "3" [label="anira::InferenceConfig" tooltip="anira::InferenceConfig"] "1" [label="anira::LibtorchProcessor" tooltip="anira::LibtorchProcessor" fillcolor="#BFBFBF"] "7" [label="anira::LibtorchProcessor::Instance" tooltip="anira::LibtorchProcessor::Instance"] "5" [label="anira::ModelData" tooltip="anira::ModelData"] "4" [label="anira::ProcessingSpec" tooltip="anira::ProcessingSpec"] "6" [label="anira::TensorShape" tooltip="anira::TensorShape"] "2" -> "3" [dir=forward tooltip="usage"] "3" -> "4" [dir=forward tooltip="usage"] "3" -> "5" [dir=forward tooltip="usage"] "3" -> "6" [dir=forward tooltip="usage"] "1" -> "2" [dir=forward tooltip="public-inheritance"] "1" -> "7" [dir=forward tooltip="usage"] "7" -> "3" [dir=forward tooltip="usage"] "7" -> "8" [dir=forward tooltip="usage"] }

LibTorch-based neural network inference processor.

The LibtorchProcessor class provides neural network inference capabilities using Facebook’s PyTorch C++ API (LibTorch). It supports loading TorchScript models and performing real-time inference with parallel processing capabilities.

See also

BackendBase, LibtorchProcessor::Instance, InferenceConfig, ModelData, SessionElement

Warning

This class is only available when compiled with USE_LIBTORCH defined

Public Functions

LibtorchProcessor(InferenceConfig &inference_config)#

Constructs a LibTorch processor with the given inference configuration.

Initializes the LibTorch processor and creates the necessary number of parallel processing instances based on the configuration’s num_parallel_processors setting.

Model Loading:

The constructor attempts to load the TorchScript model specified in the configuration. If a model function is specified, it will be used; otherwise, the default forward method is called.

Parameters:

inference_config – Reference to inference configuration containing model path, tensor shapes, and processing parameters

~LibtorchProcessor() override#

Destructor that properly cleans up LibTorch resources.

Ensures proper cleanup of all LibTorch modules, tensors, and allocated memory. All processing instances are safely destroyed.

virtual void prepare() override#

Prepares all LibTorch instances for inference operations.

Loads the TorchScript model into all parallel processing instances, allocates input/output tensors, and performs warm-up inferences if specified in the configuration.

virtual void process(std::vector<BufferF> &input, std::vector<BufferF> &output, std::shared_ptr<SessionElement> session) override#

Processes input buffers through the LibTorch model.

Performs neural network inference using LibTorch, converting audio buffers to PyTorch tensors, executing the model, and converting results back to audio buffers.

Parameters:

  • input – Vector of input buffers containing audio samples or parameter data

  • output – Vector of output buffers to receive processed results

  • session – Shared pointer to session element providing thread-safe instance access