gen_dl_pruned_modelT_gen_dl_pruned_modelGenDlPrunedModelGenDlPrunedModelgen_dl_pruned_model (Operator)

Name

gen_dl_pruned_modelT_gen_dl_pruned_modelGenDlPrunedModelGenDlPrunedModelgen_dl_pruned_model — Prune a deep learning model.

Signature

gen_dl_pruned_model( : : DLModelHandleToPrune, DLPruningHandle : DLModelHandlePruned)

Description

The operator gen_dl_pruned_modelgen_dl_pruned_modelGenDlPrunedModelGenDlPrunedModelGenDlPrunedModelgen_dl_pruned_model copies the input model DLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPrunedlmodel_handle_to_prune and returns the pruned copy in DLModelHandlePrunedDLModelHandlePrunedDLModelHandlePrunedDLModelHandlePrunedDLModelHandlePruneddlmodel_handle_pruned. Note that gen_dl_pruned_modelgen_dl_pruned_modelGenDlPrunedModelGenDlPrunedModelGenDlPrunedModelgen_dl_pruned_model supports only deep learning models of 'type'"type""type""type""type""type"='classification'"classification""classification""classification""classification""classification".

The parameter DLPruningHandleDLPruningHandleDLPruningHandleDLPruningHandleDLPruningHandledlpruning_handle provides the necessary information as e.g., the selected pruning mode, the percentage or the kernel scores. See create_dl_pruningcreate_dl_pruningCreateDlPruningCreateDlPruningCreateDlPruningcreate_dl_pruning for further information about implemented pruning modes.

Behavior in special cases:

In case the set amount of pruning 'percentage'"percentage""percentage""percentage""percentage""percentage" is too small to result in any removals, the returned model DLModelHandlePrunedDLModelHandlePrunedDLModelHandlePrunedDLModelHandlePrunedDLModelHandlePruneddlmodel_handle_pruned is simply a copy of the input model DLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPrunedlmodel_handle_to_prune.
In case the model could not be pruned successfully, an empty handle is returned.

Execution Information

Multithreading type: reentrant (runs in parallel with non-exclusive operators).
Multithreading scope: global (may be called from any thread).
Processed without parallelization.

Parameters

DLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPruneDLModelHandleToPrunedlmodel_handle_to_prune (input_control) dl_model → (handle)

Input model.

DLPruningHandleDLPruningHandleDLPruningHandleDLPruningHandleDLPruningHandledlpruning_handle (input_control) dl_pruning → (handle)

Pruning data handle.

DLModelHandlePrunedDLModelHandlePrunedDLModelHandlePrunedDLModelHandlePrunedDLModelHandlePruneddlmodel_handle_pruned (output_control) dl_model → (handle)

Pruned model.

Possible Predecessors

add_dl_pruning_batchadd_dl_pruning_batchAddDlPruningBatchAddDlPruningBatchAddDlPruningBatchadd_dl_pruning_batch, set_dl_pruning_paramset_dl_pruning_paramSetDlPruningParamSetDlPruningParamSetDlPruningParamset_dl_pruning_param

Module

Deep Learning Training

Operators