`neural_compressor.config`¶

Configs for Neural Compressor.

Module Contents¶

Classes¶

`Options`	Option Class for configs.
`BenchmarkConfig`	Config Class for Benchmark.
`AccuracyCriterion`	Class of Accuracy Criterion.
`TuningCriterion`	Class for Tuning Criterion.
`PostTrainingQuantConfig`	Config Class for Post Training Quantization.
`QuantizationAwareTrainingConfig`	Config Class for Quantization Aware Training.
`WeightPruningConfig`	Similiar to torch optimizer's interface.
`KnowledgeDistillationLossConfig`	Config Class for Knowledge Distillation Loss.
`IntermediateLayersKnowledgeDistillationLossConfig`	Config Class for Intermediate Layers Knowledge Distillation Loss.
`SelfKnowledgeDistillationLossConfig`	Config Class for Self Knowledge Distillation Loss.
`DistillationConfig`	Config of distillation.
`MixedPrecisionConfig`	Config Class for MixedPrecision.
`ExportConfig`	Config Class for Export.
`ONNXQlinear2QDQConfig`	Config Class for ONNXQlinear2QDQ.
`Torch2ONNXConfig`	Config Class for Torch2ONNX.
`TF2ONNXConfig`	Config Class for TF2ONNX.

Functions¶

check_value(name, src, supported_type[, supported_value])

Check if the given object is the given supported type and in the given supported value.

neural_compressor.config.check_value(name, src, supported_type, supported_value=[])¶: Check if the given object is the given supported type and in the given supported value.

class neural_compressor.config.Options(random_seed=1978, workspace=default_workspace, resume_from=None, tensorboard=False)¶

Option Class for configs.

property random_seed¶: Get random seed.

property workspace¶: Get workspace.

property resume_from¶: Get resume_from.

property tensorboard¶: Get tensorboard.

class neural_compressor.config.BenchmarkConfig(inputs=[], outputs=[], backend='default', warmup=5, iteration=-1, cores_per_instance=None, num_of_instance=None, inter_num_of_threads=None, intra_num_of_threads=None)¶

Config Class for Benchmark.

property backend¶: Get backend.

property outputs¶: Get outputs.

property inputs¶: Get inputs.

property warmup¶: Get warmup.

property iteration¶: Get iteration.

property cores_per_instance¶: Get cores_per_instance.

property num_of_instance¶: Get num_of_instance.

property inter_num_of_threads¶: Get inter_num_of_threads.

property intra_num_of_threads¶: Get intra_num_of_threads.

class neural_compressor.config.AccuracyCriterion(higher_is_better=True, criterion='relative', tolerable_loss=0.01)¶

Class of Accuracy Criterion.

property higher_is_better¶: Get higher_is_better.

property relative¶: Get tolerable_loss when criterion is relative.

property absolute¶: Get tolerable_loss when criterion is absolute.

property criterion¶: Get criterion.

property tolerable_loss¶: Get tolerable_loss.

class neural_compressor.config.TuningCriterion(strategy='basic', strategy_kwargs=None, timeout=0, max_trials=100, objective='performance')¶

Class for Tuning Criterion.

property max_trials¶: Get max_trials.

property timeout¶: Get timeout.

property objective¶: Get objective.

property strategy¶: Get strategy.

property strategy_kwargs¶: Get strategy_kwargs.

class neural_compressor.config.PostTrainingQuantConfig(device='cpu', backend='default', quant_format='default', inputs=[], outputs=[], approach='static', calibration_sampling_size=[100], op_type_list=None, op_name_list=None, reduce_range=None, excluded_precisions=[], quant_level=1, tuning_criterion=tuning_criterion, accuracy_criterion=accuracy_criterion)¶

Bases: _BaseQuantizationConfig

Config Class for Post Training Quantization.

property approach¶: Get approach.

property tuning_criterion¶: Get tuning_criterion.

class neural_compressor.config.QuantizationAwareTrainingConfig(device='cpu', backend='default', inputs=[], outputs=[], op_type_list=None, op_name_list=None, reduce_range=None, excluded_precisions=[], quant_level=1)¶

Bases: _BaseQuantizationConfig

Config Class for Quantization Aware Training.

property approach¶: Get approach.

class neural_compressor.config.WeightPruningConfig(pruning_configs=[{}], target_sparsity=0.9, pruning_type='snip_momentum', pattern='4x1', op_names=[], excluded_op_names=[], start_step=0, end_step=0, pruning_scope='global', pruning_frequency=1, min_sparsity_ratio_per_op=0.0, max_sparsity_ratio_per_op=0.98, sparsity_decay_type='exp', pruning_op_types=['Conv', 'Linear'], **kwargs)¶

Similiar to torch optimizer’s interface.

property weight_compression¶: Get weight_compression.

class neural_compressor.config.KnowledgeDistillationLossConfig(temperature=1.0, loss_types=['CE', 'CE'], loss_weights=[0.5, 0.5])¶: Config Class for Knowledge Distillation Loss.

class neural_compressor.config.IntermediateLayersKnowledgeDistillationLossConfig(layer_mappings=[], loss_types=[], loss_weights=[], add_origin_loss=False)¶: Config Class for Intermediate Layers Knowledge Distillation Loss.

class neural_compressor.config.SelfKnowledgeDistillationLossConfig(layer_mappings=[], temperature=1.0, loss_types=[], loss_weights=[], add_origin_loss=False)¶: Config Class for Self Knowledge Distillation Loss.

class neural_compressor.config.DistillationConfig(teacher_model=None, criterion=criterion, optimizer={'SGD': {'learning_rate': 0.0001}})¶

Config of distillation.

Parameters:

teacher_model (Callable) – Teacher model for distillation. Defaults to None.
features (optional) – Teacher features for distillation, features and teacher_model are alternative. Defaults to None.
criterion (Callable, optional) – Distillation loss configure.
optimizer (dictionary, optional) – Optimizer configure.

property criterion¶: Get criterion.

property optimizer¶: Get optimizer.

property teacher_model¶: Get teacher_model.

class neural_compressor.config.MixedPrecisionConfig(device='cpu', backend='default', inputs=[], outputs=[], tuning_criterion=tuning_criterion, accuracy_criterion=accuracy_criterion, excluded_precisions=[])¶

Bases: PostTrainingQuantConfig

Config Class for MixedPrecision.

class neural_compressor.config.ExportConfig(dtype='int8', opset_version=14, quant_format='QDQ', example_inputs=None, input_names=None, output_names=None, dynamic_axes=None)¶

Config Class for Export.

property dtype¶: Get dtype.

property opset_version¶: Get opset_version.

property quant_format¶: Get quant_format.

property example_inputs¶: Get example_inputs.

property input_names¶: Get input_names.

property output_names¶: Get output_names.

property dynamic_axes¶: Get dynamic_axes.

class neural_compressor.config.ONNXQlinear2QDQConfig¶: Config Class for ONNXQlinear2QDQ.

class neural_compressor.config.Torch2ONNXConfig(dtype='int8', opset_version=14, quant_format='QDQ', example_inputs=None, input_names=None, output_names=None, dynamic_axes=None, recipe='QDQ_OP_FP32_BIAS', **kwargs)¶

Bases: ExportConfig

Config Class for Torch2ONNX.

class neural_compressor.config.TF2ONNXConfig(dtype='int8', opset_version=14, quant_format='QDQ', example_inputs=None, input_names=None, output_names=None, dynamic_axes=None, **kwargs)¶

Bases: ExportConfig

Config Class for TF2ONNX.

neural_compressor.config¶

Module Contents¶

Classes¶

Functions¶

`neural_compressor.config`¶