Set up benchmarking arguments for trt

Returns the underlying string object. Returns: str

Ceiling implementation. \param arg value to round \return rounded value

\brief Create a builder configuration object. \see IBuilderConfig

\brief Create an instance of an IRuntime class. \param logger The logging class for the runtime.

Exponential function. \param arg function argument \return e raised to \a arg template<typename T> typename enable<expr,T>::type exp(T arg) { return f

Logs a message to stdout with FINISH severity. Args: message (Union[str, Callable() -> str]): A stri

Creates a generator that yields ``Tuple[str, List[IterationResult]]`` - runner names and corresponding outputs.

Load amax from calibrator. Updates the amax buffer with value computed by the calibrator, creating it if necessary. *args and **kwarg

Dummy network with resnet-like architecture.

Resizes or reshapes the array to the specified shape. If the allocated memory region is already large enough, no reallocatio

\brief Set the cuda stream that is used to profile this network. \param stream The cuda stream used for profiling by the builder. \see getProfileStr

Square root. \param arg function argument \return square root of \a arg template<typename T> typename enable<expr,T>::type sqrt(T arg) { return functi

\brief Add an elementwise layer to the network. \param input1 The first input tensor to the layer. \param input2 The second input tensor to the layer

\brief Builds and serializes a network for the given INetworkDefinition and IBuilderConfig. This function allows building and serialization of a netw

Copies from this device buffer to the provided host buffer. Args: host_buffer (numpy.ndarray): The h

\brief Create a network definition object Creates a network definition object with immutable properties specified using the flags parameter. CreateNe

\brief Enqueue inference on a stream. This method requires an array of input and output buffers. The mapping from tensor names to indices can be quer

Returns information about the inputs of the model. Shapes here may include dynamic dimensions, represented by ``None``. Must

Loads an ONNX model according to arguments provided on the command-line. Returns: onnx.ModelProto: The model that was lo

Saves an ONNX model according to arguments provided on the command-line. Args: model (onnx.ModelProto): The ONNX model t

\brief Add a constant layer to the network. \param dimensions The dimensions of the constant. \param weights The constant value, represented as weigh

Return the absolute max of all tensors collected

\brief Create an instance of an IBuilder class. \param logger The logging class for the builder. unnamed namespace avoids linkage surprises when lin

Returns a specified default value if the provided value is None. Args: value : The value. default : The default value to use

\brief Retrieve the name corresponding to a binding index. This is the reverse mapping to that provided by getBindingIndex(). For optimization profi

\brief Get the kernel weights of the convolution. \see setKernelWeights()

\brief Get the number of binding indices. There are separate binding indices for each optimization profile. This method returns the total over all pr

Imports a Graph from an ONNX Graph. Args: onnx_graph (onnx.GraphProto): The ONNX graph to import. tensor_ma

Natural logorithm. \param arg function argument \return logarithm of \a arg to base e template<typename T> typename enable<expr,T>::type log(T arg) {

Natural logorithm. \param arg function argument \return logarithm of \a arg to base e template<typename T> typename enable<expr,T>::type log(T arg) {

Measures the time it takes to run Pythonic inference code. Statement given should be the actual model inference like forward() in torch.

Block with BN+ReLU

Runs this tool with the specified arguments. This must be called only after calling `parse()`. Args: args (Names

Save the ONNX model to the given location. :param output_path: Path pointing to the location where to write out the updated ONNX mode

Return the sigmoid of the input.

Synchronizes the stream.

Returns the global Polygraphy CUDA wrapper. Returns: Cuda: The global CUDA wrapper.

Returns a class (not an instance) that will provide a dictionary-like interface with runtime type checks. Note: The types are provided l

Allocate binding buffers for trt based on provided types and shapes dict

Converts current model into an TRT engine. Uses provided converter to convert object or default_convert is used instead if available.

Deactivate the runner. For example, this may involve freeing CPU or GPU memory. Generally, you should use a context manager instead

\brief Define a test for logging \param[in] name The name of the test. This should be a string starting with "TensorRT" and containing dot-separated

Absolute value. \param arg operand \return absolute value of \a arg template<typename T> typename enable<T,T>::type fabs(T arg) { return unary_special

Absolute value. \param arg operand \return absolute value of \a arg template<typename T> typename enable<T,T>::type fabs(T arg) { return unary_special

\brief Get the bias weights for the convolution. \see setBiasWeights()

Creates a data loader according to arguments provided on the command-line. Returns: Sequence[OrderedDict[str, numpy.ndar

Returns an inline safe string if the provided string is an identifier and raises an exception otherwise. Args: ident (str): The stri

Loads a TensorRT Network model according to arguments provided on the command-line. Returns: tensorrt.INetworkDefinition

\brief Mark a tensor as a network output. \param tensor The tensor to mark as an output tensor. \warning It is an error to mark a network input as a