Guesses the data format of a given shape. Args: shape (Tuple[int]): The shape, including batch dimension. Retur

\brief Enqueue inference on a stream. \param stream A cuda stream on which the inference kernels will be enqueued. \return True if the kernels were

Create a directory to save test logs Args: dir_path (str): directory to create / clean. do_clean_dir (bool): boolean indicating w

Error function. \param arg function argument \return error function value of \a arg template<typename T> typename enable<expr,T>::type erf(T arg) { re

Error function. \param arg function argument \return error function value of \a arg template<typename T> typename enable<expr,T>::type erf(T arg) { re

Export an onnx-graphsurgeon Graph to an ONNX GraphProto. Args: graph (Graph): The graph to export. do_type_

Run inference with ONNX-Runtime. This can be used to retrieve values/shapes/data types for all tensors in the model when oth

Frees the GPU memory associated with this array. You can also use a context manager to ensure that memory is freed. For example:

\brief Return the number of components per element. This is always 1 for non-vectorized formats. \return the number of components per element.

\brief Return DataType of the input/output of algorithm. \return the data type.

\brief Get the enqueueEmitsProfile state. \return The enqueueEmitsProfile state. \see IExecutionContext::setEnqueueEmitsProfile()

\brief Return implementation of the algorithm.

\brief Get the scale mode. \see setMode()

\brief Get number of optimization profiles. This is one higher than the index of the last optimization profile that has be defined (or zero, if none

\brief Get the padding. \see setPostPadding()

\brief Get the pre-padding. \see setPrePadding()

\brief Get the minimum / optimum / maximum dimensions for an input tensor given its name under an optimization profile. \param tensorName The name of

\brief Get the profiler. \see IProfiler setProfiler()

\brief Get the minimum / optimum / maximum values for an input shape tensor. If the shape values have not been set previously with setShapeValues(),

\brief Get the stride for the output slice. \return The slicing stride, or an invalid Dims structure. If the fourth input is present and non-null, t

\brief Return tactic of the algorithm.

\brief Return the index of the vectorized dimension or -1 for non-vectorized formats. \return the index of the vectorized dimension.

Retrieve the batches. This is a generator object, so you can use it within a loop as: for batch, images in batcher.get_batch():

Retrieve the batches. This is a generator object, so you can use it within a loop as: for batch, images in batcher.get_batch():

Retrieve the batches. This is a generator object, so you can use it within a loop as: for batch, images in batcher.get_batch():

Return a list of the inputs bindings and a list of the output bindings

Generates the directory name to save all files relevant to the model's quantization. Args: hyperparams (Dict): dictionary with neces

Gets all zeros frames to use as first decoder input PARAMS ------ memory: decoder outputs RETURNS -------

Provided helper function to obtain a profile optimization.

* getopt_internal -- * Parse argc/argv argument vector. Called by user level routines. */

Identity block with no shortcut convolution

Indents the provided block of text. Args: block (str): The text to indent. level (int): The number of tabs to indent with.

Execute inference on a batch of images. :param batch: A numpy array holding the image batch. :return A list of outputs as num

Execute inference on a batch of images. The images should already be batched and preprocessed, as prepared by the ImageBatcher class.

\brief Runs the TensorRT inference engine for this sample. \details This function is the main execution function of the sample. It allocates the buff

Initializes attention rnn states, decoder rnn states, attention weights, attention cumulative weights, attention context, stores memory

Get the specs for the input tensor of the network. Useful to prepare memory allocations. :return: Two items, the shape of the input t

Checks if parent heritage follows a specific 'pattern'. Args: parent_info (dict): dictionary with parent's information. pattern (

Checks if 'parent_class' is of type 'type'. Examples of types: Conv, BatchNorm, Dropout, Activation.

Whether this data loader will randomly generate data rather than use real data. Returns: bool

This method provides the overall functionality for iteratively debugging. It essentially does: :: while True:

ideal latency

Build and load engines for TensorRT accelerated inference. Export ONNX models first, if applicable. Args: engine

Loads a TensorFlow graph according to arguments provided on the command-line. Returns: tf.Graph

Loads the NNSemanticCheckpoint instance from checkpoint.toml file.

Loads `tfrecord` dataset and calculates the data's mean, min, and max values. Args: model_name (str): model name for data pre-proces

Nearest integer. \param arg half to round \return nearest integer, rounded away from zero in half-way cases template<typename T> typename enable<long,