* Initialize & check a TensorList based on an input shape * Allocate it as float */

to be called in destructor

Associates the task with a scheduler and sets the state to Pending. */

* @brief Check if `out_shape` matches the number of channels of `in_shape` and spatial dimensions * of `output_sizes` as if it was generated with G

* @brief Wrap tensor data buffer into nvcv::Image. * * @param data pointer to beginning of the buffer * @param shape underlying tensor shape * @pa

Checks whether using this `DataNode` in a CPU operator is legal. The function checks whether it's legal to pass it as an input to a CPU opera

Create a temp Caffe2 LMDB dir with one key-value. Returns path to the dir.

Returns the default number of exponent bits for a given number of bits.

Determines the variables affected inside a control flow statement.

Returns the pipeline's state as a serialized Protobuf string. Also, allows to pass additional data to be saved in the checkpoint.

Get the numpydoc-formatted docstring section for keywords arguments. schema : OpSchema Schema of the operator to be documented

Compute learning rate at given step.

Get the Parameter representing positional inputs at `idx`. Automatically mark it as optional. The DataNode annotation currently hides the possibil

Private method is used to execute operator from a Compose pipeline Note: Do not call directly this is to unify the architecture with

Checks whether x is a generator function or a callable object where __call__ is a generator function

Tests whether the function or method is an instance of a known type.

Custom wrapper over optree.tree_map that filters it out from the user-visible stack trace for error reporting purposes.

Apply 1x1 conv to change layer width if necessary.

Return a proper display name of operator based on the API it was instantiated in. Only valid after `__init__` kwargs were split into

Process arguments: validate, deprecate and add to spec, handling appropriate data marshalling.

Based on the schema name (for example "Resize" or "experimental__readers__Video") transform it into Python-compatible module & operator name infor

Checks if remove any padded sample and how much. Calculates the number of padded samples in the batch for each pipeline wrap

Unify the arguments (or argument inputs) between __call__ and __init__ - ops API only. If the same argument was passed in both invocation it shou

Runs the reader and obtains one result (batch or sample, depending on `batch_size`). Do not call this function directly. Use `__call

Run the operator and store the result in self._results