* Check, if the "depleted" trace exists. */

* Verify the value of "depleted" trace. * @param shall_be_depleted Expected value. */

Wraps a constant value which can then be used in :meth:`nvidia.dali.Pipeline.define_graph` pipeline definition step. If the `value` argument

* @brief Add a window element, that has `NumLanes()` consecutive lanes and starting at `input`, * to the buffer. */

* @brief Helper method that serialized OpSpec * decouples spec class from dali_proto */

* * @brief For tensor T of shape (s_0, s_1, ..., s_{n-1}) returns a n-1 dimensional tensor T' * of shape (s_1, s_2, ..., s_{n-1}), such t

* @brief Update current state obeying the enforced state. * BatchContiguity::Automatic is always allowed and does not change the state *

Helper for indenting the log messages to resemble visited scopes

Context manager: build a MXNet RecordIO pipeline using 3 split .rec files. The original .idx file is reused unchanged – it stores global offsets

Context manager: build a 3-shard TFRecord pipeline (same data file, split index). Using three logical files ensures file-level shuffling produces

Clears the predicate set efficiently

Please refer to the ``RandomCrop`` operator for more details.

* @return true, if this box is empty (its volume is 0) */

Feeds input data to DALI pipeline. Supports multiple inputs. :param dali_pipeline: DALI pipeline object :param input_data: Input data to

Freezes this scope.

Returns a :py:class:`DType` associated with given :py:class:`nvidia.dali.types.DALIDataType`.

Loads a state_dict created by an earlier call to state_dict. Untested.