Package 'churon'

S3 Methods for TensorInfo and RSession

Description

These methods provide convenient access to object properties.

Usage

## S3 method for class 'TensorInfo'
x$name

Arguments

x	A TensorInfo or RSession object
name	Property name to access

Value

For name and data_type, a character scalar identifying the tensor or its element type. For shape, an integer vector of tensor dimensions (negative dimensions indicate dynamic sizes). For a method name, a function bound to the object; for an unknown name, NULL.

---

Optimize Session Performance Batch Process Data

Description

Process data in batches for memory efficiency with large datasets.

Usage

batch_process_data(session, data_list, batch_size = 32)

Arguments

session	An RSession object created by onnx_session()
data_list	A list of input data to process
batch_size	Number of items to process in each batch

Value

A list of results from batch processing

---

Check if ONNX Runtime is Available

Description

This function checks if ONNX Runtime is properly configured and available.

Usage

check_onnx_runtime_available()

Value

Logical indicating whether ONNX Runtime is available

---

Get ONNX Runtime Information

Description

This function returns information about the current ONNX Runtime configuration.

Usage

get_onnx_runtime_info()

Value

A list containing ONNX Runtime configuration information

---

Install ONNX Runtime

Description

Download and install ONNX Runtime library for your platform. This is required before using the churon package if ONNX Runtime is not already installed on your system.

Usage

install_onnx_runtime(version = "1.29.0", quiet = FALSE, destdir, ...)

Arguments

version	Character string specifying the ONNX Runtime version to install. Defaults to "1.29.0". Supported versions are "1.28.0" and "1.29.0"; "latest" selects "1.29.0".
quiet	Logical. If TRUE, suppress download progress messages.
destdir	Required character string naming the installation directory.
...	Additional arguments passed to download.file()

Details

This function installs external software only when explicitly called. Supply destdir yourself; no installation directory is selected by default. The library is placed in destdir/lib and configured for this R session. In later sessions, set ORT_DYLIB_PATH to the installed library file before loading churon. Restart R before switching an already loaded runtime.

Value

Invisible TRUE on success, stops with error on failure.

---

ONNX Example Models

Description

List available example models bundled with the package.

Usage

onnx_example_models()

Value

A named character vector with model names as names and paths as values

---

ONNX Example Session

Description

Create a session with an example model.

Usage

onnx_example_session(model_name = "mnist", providers = NULL)

Arguments

model_name	Character string specifying the model name (default: "mnist")
providers	Optional execution providers

Value

An RSession object

---

Get Input Information

Description

Retrieve information about model input tensors.

Usage

onnx_input_info(session)

Arguments

session

An RSession object created by onnx_session()

Value

A list of TensorInfo objects containing input tensor metadata

Examples

if (check_onnx_runtime_available()) {
  session <- onnx_example_session()
  input_info <- onnx_input_info(session)
  print(input_info)
}

---

Get Model Path

Description

Get the model path from a session.

Usage

onnx_model_path(session)

Arguments

session

An RSession object created by onnx_session()

Value

Character string with the model path

Examples

if (check_onnx_runtime_available()) {
  session <- onnx_example_session()
  model_path <- onnx_model_path(session)
  cat("Model path:", model_path, "\n")
}

---

Get Output Information

Description

Retrieve information about model output tensors.

Usage

onnx_output_info(session)

Arguments

session

An RSession object created by onnx_session()

Value

A list of TensorInfo objects containing output tensor metadata

Examples

if (check_onnx_runtime_available()) {
  session <- onnx_example_session()
  output_info <- onnx_output_info(session)
  print(output_info)
}

---

Get Execution Providers

Description

Get the execution providers available for the session.

Usage

onnx_providers(session)

Arguments

session

An RSession object created by onnx_session()

Value

A character vector of available execution providers

Examples

if (check_onnx_runtime_available()) {
  session <- onnx_example_session()
  providers <- onnx_providers(session)
  cat("Available execution providers:", paste(providers, collapse = ", "), "\n")
}

---

Run ONNX Inference

Description

Execute inference on an ONNX model with input data.

Usage

onnx_run(session, inputs)

Arguments

session	An RSession object created by onnx_session()
inputs	A named list of input tensors. Names should match model input names.

Value

A named list of output tensors

Examples

if (check_onnx_runtime_available()) {
  session <- onnx_example_session()
  inputs <- list(Input3 = array(0, dim = c(1, 1, 28, 28)))
  outputs <- onnx_run(session, inputs)
}

---

Create ONNX Session

Description

Create a new ONNX Runtime session from a model file.

Usage

onnx_session(model_path, providers = NULL)

Arguments

model_path	Character string specifying the path to the ONNX model file
providers	Optional character vector specifying execution providers to use. Available providers: "cuda", "tensorrt", "directml", "onednn", "coreml", "cpu". If NULL, uses default provider priority.

Value

An RSession object for running inference

Examples

if (check_onnx_runtime_available()) {
  # Create session with default providers
  session <- onnx_example_session()

  # Create session with specific providers
  session <- onnx_example_session(providers = "cpu")
}

---

Safe ONNX Run

Description

Run inference with automatic error handling and monitoring.

Usage

safe_onnx_run(session, inputs, monitor_performance = FALSE)

Arguments

session	An RSession object created by onnx_session()
inputs	A named list of input tensors
monitor_performance	Logical indicating whether to monitor performance

Value

Result of inference or NULL if failed

---

Safe ONNX Session Creation

Description

Safe ONNX Session Creation

Usage

safe_onnx_session(model_path, providers = NULL)

Arguments

model_path	Character string specifying the path to the ONNX model file
providers	Optional character vector specifying execution providers

Details

Create an ONNX session with automatic error handling.

Value

An RSession object or NULL if creation fails

---